The Gulf of Cádiz (Figure 1) is located in the NE Atlantic Ocean, west of the Gibraltar Arc between the Iberian Peninsula and the Moroccan margin. The area has experienced a complex geological history related to plate tectonic interactions which have resulted in several episodes of rifting, compression and strike-slip motion due to the closure of the Tethys Ocean, the opening of the N Atlantic, and the African-Eurasian convergence since the Cenozoic (Maldonado et al., 1999). Due to this ongoing compression, rapidly deposited sediments have been dewatered progressively and formed many MVs and fluid escape structures (Díaz-del-Rio et al., 2003). Since their discovery in 1999 (Gardner, 2001) more than 60 MVs have now been identified (León et al., 2012), ranging from 800 to 3,500 m in diameter and some as high as 300 m above the seafloor. Most MVs in the area have been built up from episodes of mud-breccia flows (Medialdea et al., 2009) and show evidence of gas saturation, methane hydrate, hydrogen sulphide and the presence of chemosynthetic fauna (e.g., Somoza et al., 2003; Niemann et al., 2006a; Rodrigues et al., 2010, 2011).

Sediment cores were collected from seven mud volcanoes (Porto, Bonjardim, Carlos Ribeiro, Captain Arutyunov, Darwin, Meknes and Mercator MV) from the Gulf of Cádiz (Figure 1) during the EU HERMES project expedition cruises MSM1-3 on the Research Vessel Maria S. Merian (12^th April – 19^th May 2006) and JC10 (Leg 1) on the Royal Research Ship James Cook (14^th May to 2^nd June 2007) using a range of sampling methods and devices (Table 1). Additional sediment cores were also taken during Expedition JC10 from reference seafloor sites >1 km away from each MV sampled.

On board ship, sediment cores from each MV site were sampled aseptically at different depths (ranging from 0.01 to 5.25 mbsf; meters below seafloor; Table 1) using sterile pre-cut (luer-end removed) 50 ml volume syringes. All syringe mini-cores were then capped and sealed inside gas-tight aluminum bags under anoxic conditions with a nitrogen atmosphere and an Anaerocult A (Merck; Cragg et al., 1992), and stored at 4°C or frozen directly at −80°C. On board ship, further sediment samples were taken for cell counts, sediment pore water chemistry, gas analysis and activity measurements (only cruise JC-10). All mini-cores, samples for AODC and chemical analyses were then transported back to the UK for processing. Note 4°C stored samples were used for methanogen enrichments and chemical analysis, and −80°C samples for molecular analysis.

Samples for AODC were fixed on board ship in filter-sterilized (0.2 μm) 4% (w/v) formaldehyde as described by Fry (1988). Preserved samples were then stained [0.1% (w/v) acridine orange] and counted on black polycarbonate membrane filters using a Zeiss Axioskop epifluorescence microscope.

Sediment for methane gas analysis was transferred on board ship into 20 ml volume serum vials with 10 ml of 10% (w/v) KCl, sealed and stored for equilibration. Headspace gas was analyzed using a Perkin Elmer/Arnel Clarus 500 natural gas analyzer with a flame ionization detector and a thermal conductivity detector. Methane and carbon dioxide concentrations were determined using a calibration with standard gases (Scott Specialty Gases). Pore waters were obtained from sediments on board ship using a pore-water squeezer as described (Wellsbury et al., 2002). Anions (including sulphate and acetate concentrations) were determined using an ICS-2000 Ion Chromatography System with two Ionpac AS15 columns in series and an Anion Self-Regenerating Suppressor (ASRS-ULTRA II 4 mm) in combination with a DS6 heated conductivity cell (Dionex UK Ltd.) as described (Webster et al., 2009). Cations (including methylamines) were analyzed using a DX-120 ion chromatograph (Dionex UK Ltd.) fitted with an IonPac CS16 column, a CSRS 300 4-mm suppressor, and a conductivity detector and methanesulfonic acid eluent (32 mM) at a flow rate of 0.75 ml min⁻¹ (Watkins et al., 2012).

Radiotracer experiments using ¹⁴C-labelled substrates were carried out as previously described (Parkes et al., 2007, 2010). Briefly, sediment was subsampled using either Perspex mini-cores (diameter 2.0 cm, length 20 cm) with silicone-filled side injection holes (4 μl substrate per 1 cm depth interval) or 5-ml sterile pre-cut (luer-end removed) syringes closed with rubber stoppers (7.5 μl substrate per syringe). Samples were then preincubated in nitrogen-flushed gas-tight aluminum bags at 10°C to equilibrate for 24 h before radiotracer injection. Samples were injected with one of four ¹⁴C-substrates in batches of four (one control and three measurements) using either ¹⁴C-sodium bicarbonate (9.90 KBq μl⁻¹), ¹⁴C-sodium acetate (6.60 KBq μl⁻¹), ¹⁴C-methanol (2.05 KBq μl⁻¹) or ¹⁴C-methylamine (7.36 KBq μl⁻¹) and incubated at in situ bottom water temperature for 7 to 24 h in nitrogen-flushed gas-tight aluminum bags. Microbial reactions were stopped by transferring the incubated sediment into 30 ml glass serum vials containing 7 ml of 1 M NaOH and sealed. All samples were stored inverted at room temperature prior to processing. For ¹⁴C-CH₄ analysis, samples were magnetically stirred whilst the headspace gas was flushed for 20 min with 5% O₂: 95% N₂ and passed over copper oxide at 900°C to convert ¹⁴C-CH₄ to ¹⁴C-CO₂. Flushed gases were bubbled through a series of three scintillation vials of 10 ml of Hi-Safe 3 scintillation cocktail (Canberra-Packard) containing 7% (v/v) ß-phenethylamine to absorb any ¹⁴C-CO₂. Scintillation vials were counted in a scintillation counter and label turnover rates and potential activity rates calculated as described (Parkes et al., 2010). Methane production rates were calculated based on the proportion of labelled gas produced from the ¹⁴C-substrate, the dissolved pore water substrate or total CO₂ concentration adjusted for sediment porosity, and the incubation time. For methylamine methanogenesis, rates were obtained by multiplying the substrate turnover value by an assumed methylamine concentration of 5 μM (as methylamine concentration in pore water was consistently below detection limit of approx. 5 μM), as previously used for measuring methylamine methanogenesis in Gulf of Cádiz MVs (Maignien et al., 2013). Therefore, methylamine methanogenesis rates should be considered as maximum potential rates assuming in situ concentration was the detection limit.

Cold-stored (4°C) sediment mini-cores from each MV were then further sub-sampled under aseptic conditions in an anaerobic chamber and used as inoculum for sediment slurry enrichments. Replicate slurries (50 ml) were made up using 25% (v/v) sediment (see Table 1 for sediment depths used) in artificial seawater (Köpke et al., 2005) contained in 100 ml serum bottles and supplemented with the substrates 12 mM sodium acetate, 1.5 mM benzoic acid, 4 mM hexadecane, 5 mM methanol, or 5 mM methylamine. Additionally, one set of sediment slurries were left without supplements as an unamended control and another set of slurries had their headspace gas replaced with H₂/CO₂ (80:20) as substrates. All slurries, except those with H₂/CO₂ had their headspace gas replaced with N₂/CO₂ and all methanogen enrichments were then incubated at 10°C in the dark for up to 170 days (cruise MSM1-3) or 300 days (cruise JC10).

At several time points during the incubation, samples of headspace gas (2 ml) were removed from all sediment slurries and analyzed by gas chromatography as above. For standardization, potential methanogenesis (methane production) rates were calculated from total methane produced after 130 days incubation (or in some samples after 300 days) with the methane production value from the relevant unamended control subtracted (32 out of 40 control slurries produced methane; Table 2).

Genomic DNA was extracted from methanogenic sediment slurry samples using the FastDNA® SPIN Kit for Soil (MP Biomedicals) as described (Webster et al., 2003). Essentially, 1 ml of sediment slurry was placed in a lysing matrix E tube (MP Biomedicals) and centrifuged at 15,000 × g for 1 min to pellet cells and sediment. Sediment pellets were then re-suspended in 800 μl of sodium phosphate buffer and 120 μl MT buffer (MP Biomedicals) before lysis in a FastPrep® 24 instrument (MP Biomedicals) for 2 X 30s at speed 5.5 ms⁻¹. All remaining steps were as the manufacturer’s protocol, except that some spin and incubation times were extended (see Parkes et al., 2010). DNA was eluted in 100 μl molecular grade water (Severn Biotech) and stored at −80°C until required.

Bacterial and archaeal 16S rRNA genes were amplified by direct PCR from sediment slurry DNA extracts using DreamTaq DNA polymerase (Thermo Fisher Scientific) with primer pairs 357FGC-518R or SAfGC-PARCH519R, respectively and analyzed by DGGE (see Webster et al., 2006). Additionally, methyl-coenzyme M reductase genes (mcrA) were amplified by nested PCR using primers ME1f-ME2r and MLf-MLr without a GC-clamp and analyzed on 6–12% gradient (w/v) polyacrylamide DGGE gels with a 25–50% denaturant gradient (Webster et al., 2011). All DGGE gels were stained with SYBR Gold nucleic acid stain (Invitrogen), viewed under UV and images captured using a Gene Genius Bio Imaging System (Syngene). DGGE bands were excised, re-amplified by PCR, sequenced (O’Sullivan et al., 2008) and analyzed using the NCBI nucleotide BLAST program.¹

A selected number of Gulf of Cádiz methanogen enrichments showing methane production were used as inoculum in attempts at isolating pure methanogens. Approximately 1 ml of enrichment culture was subcultured into 10 ml of artificial seawater (ASW) medium as described (Watkins et al., 2012) and supplemented with either 10 mM methylamine, 10 mM methanol or H₂/CO₂ (80,20, 0.1 MPa) as substrates. Growth was based on methane production monitored by gas chromatography. Methanogens were then isolated using a combination of deep-agar shake tubes and/or a dilution-to-extinction series with antibiotics to inhibit bacterial growth at 25°C (Parkes et al., 2010; Watkins et al., 2012). Methanogens were identified by PCR of the 16S rRNA and mcrA gene as described previously (Watkins et al., 2012).

MV sediment samples stored for molecular analysis were used to conduct a microbial diversity survey using three different sequencing approaches. DNA was extracted from 5 g of each MV sample (Table 1) using the FastDNA® SPIN Kit for Soil (MP Biomedicals) as described (Webster et al., 2003). Bacterial (V1-V5 region) and archaeal (V2-V5 region) 16S rRNA genes were amplified by PCR from sediment samples using DreamTaq DNA polymerase (Thermo Fisher Scientific Inc.) with primers 27F-907R and 109F-958R, respectively, as described (Webster et al., 2006). Three replicate PCR products for each sample were cleaned, pooled and cloned in pGEM-T Easy vector (Promega) and transformed into Escherichia coli JM109 competent cells (Promega) according to manufacturer’s protocol. Recombinant colonies were picked, grown overnight at 37°C in 96-well plates containing LB liquid medium with 7.5% (v/v) glycerol and 100 μg ml⁻¹ ampicillin, and the gene libraries stored at −80°C. Approximately 40–70 recombinant clones from each Archaea and Bacteria 16S rRNA gene library were amplified by PCR with M13 primers, checked and sequenced using an ABI 3130xl Genetic Analyzer (Applied Biosystems). Sequence chromatographs were analyzed using the Chromas Lite software package version 2.01,² and gene sequences identified using nucleotide BLAST³ against the nucleotide collection and reference RNA sequence databases. New sequences reported here have been submitted to the NCBI database under accession numbers MT825700 - MT826197.

DNA from some MV samples was also amplified using methods of the International Census of Marine Microbes (ICoMM). The V6 region of the 16S rRNA gene from Archaea and Bacteria was amplified and subjected to 454 pyrosequencing on a GS20. All PCR methods, primers and analysis tools are detailed on the ICoMM website (https://vamps2.mbl.edu; Sogin et al., 2006). Clusters were generated using the single-linkage pre-clustering algorithm to smooth sequencing errors and reduce noise, followed by primary pairwise, average linkage clustering. OTUs were created using clustering thresholds of 3%, corresponding to 97% similarity (Huse et al., 2010). Tag sequences are publicly available from ICoMM as the datasets ICM_CFU_Av6 and ICM_CFU_Bv6.

Additionally, MV DNA samples were amplified using the universal 16S rRNA gene primers (V4 region) 515F-806R (Caporaso et al., 2011) and Archaea-specific primers 519F-958R (V4-V5). Amplicons for Illumina Miseq sequencing were made by dual indexed nested PCR (D’Amore et al., 2016) designed to amplify the 16S rRNA gene and incorporate the illumina adapters and sample identification barcodes as described in the Nextera DNA sample preparation guide (Illumina Inc.). All primers were synthesised by Integrated DNA Technologies, Inc. and PCR amplifications, purification, pooling and sequencing on an Illumina MiSeq platform were carried out by the Centre for Genomic Research (CGR), University of Liverpool. Raw Fastq files were trimmed for the presence of Illumina adapter sequences using Cutadapt version 1.2.1 (Martin, 2011) and for a minimum window quality score of 20 using Sickle version 1.200 (Joshi and Fass, 2011) by CGR who provided trimmed data suitable for downstream analysis. Overlapping regions within the paired-end Illumina sequencing reads were aligned to generate contigs and quality filtered using USEARCH (Edgar, 2010). Further analysis of sequencing data was performed in QIIME (Caporaso et al., 2010). Representative OTUs were picked with pick_de_novo_otus.py command in QIIME at 97% similarity and taxonomy assigned with the Greengenes database (DeSantis et al., 2006). Singletons, unidentified and non-specific sequences (e.g., Bacteria sequences in the Archaea-specific library) were removed.

Geochemical pore water profiles (methane, sulphate, sulphide, and chloride) from seven mud volcanoes in the Gulf of Cádiz sampled during two expeditions (MSM1-3 and JC10) are shown in Figure 2. All MV sites studied differed in their geochemical profiles. The majority of MV sediments had consistent or rapid sulphate removal with depth, demonstrating active prokaryotic sulphate reduction. Sulphate was completely removed in Bonjardim, Porto, Meknes, Carlos Ribeiro, Captain Arutyunov, and Darwin MVs between ~0.2 and 1.2 meters below seafloor (mbsf; Figure 2), whereas Mercator MV had sulphate concentrations of 14.0 mM (station 019) and 26.7 mM (station 238) at the bottom of the core for each station (1.63 mbsf; Figure 2E). For the majority of MV sediments with active sulphate removal, methane concentrations increased rapidly with depth reaching saturation at atmospheric pressure by 0.5 mbsf, indicating that active methanogenesis was occurring in the subsurface (with maximum concentrations of 7.5 and 6.0 mmol CH₄ l⁻¹ sediment at Porto MV station 143 and Captain Arutyunov MV station 227, respectively; Figure 2). Despite incomplete sulphate removal at Mercator MV, high methane concentrations (e.g., 4.2 mmol CH₄ l⁻¹ sediment at station 009) were detected throughout the sediment profile for all stations analyzed. Sulphide profiles at most MV sites peaked at the interface between the methane and sulphate gradients (Figure 2), indicative of a distinct sulphate methane transition zone (SMTZ) with active anaerobic oxidation of methane (AOM). The exception being Mercator MV which had no detectable sulphide (Figure 2E).

The methane and sulphate profiles at the Darwin and Carlos Ribeiro MV reference sites (stations 025 and 045, respectively) were very different to those of the MV sediments, with slower, linear sulphate removal (14.1 and 17.1 mM at ~4.5 mbsf), and very low CH₄ concentrations in the subsurface (1–7 μmol CH₄ l⁻¹ sediment; Figure 2). Geochemical profiles at the Mercator MV reference sites 002 and 004 were similar to MV sediments with sulphate depletion and CH₄ production in the subsurface, albeit at deeper depths (>2 mbsf). This may suggest that the Mercator MV reference sites are influenced by nearby MV activity, despite pore water salinity not being elevated. Discovery of a nearby buried mud volcano may also be influencing the observed biogeochemistry at this station (Perez-Garcia et al., 2011).

Mercator MV was clearly distinct from the other MV sites due to extremely high salinity at this site (Haffert et al., 2013) with chloride concentration in the surface sediments (e.g., stations 009 and 019) being 3.5-fold higher (2,011 mM) than typical sea water values (Figure 2E). Salinity further increased with depth to 10x seawater chloride values (5,585 mM) at 2.3 mbsf to almost halite saturation levels (~ 5,800 mM) as reported previously (Maignien et al., 2013). Carlos Ribeiro MV and Meknes MV had salinity values lower than that of seawater with chloride around 200 and 300 mM, respectively, at depth. All other MV (Bonjardim, Porto, Captain Arutyunov and Darwin) and reference sites had salinities close to seawater concentrations throughout their depth profiles (Figure 2).

At all sites, AODC cell numbers at the sediment surface (0.01 to 0.20 mbsf) were within the range observed previously (0.03–6.96 ×10⁹ cells cm⁻³ sediment) for subseafloor sediments (Parkes et al., 2000) after which they decreased rapidly to 1.16 to 8.10 ×10⁶ cells cm⁻³ sediment at 0.2–0.5 mbsf (Figure 2) corresponding with the approximate depth of the SMTZ (Figure 2). Cell numbers below the SMTZ (varying depth 0.2–0.5 mbsf) then fell below the expected lower limit of the global subseafloor sediment trend line (0.95–3.89 × 10⁶ cells cm⁻³ sediment for 0.5–4.0 mbsf), where they remained relatively constant to the bottom of the sediment core. Mercator MV reference station 004 and Porto MV station 144 were however, exceptions to this observation. Cell counts at Mercator MV station 004 followed the global trend in subseafloor sediments (Figure 2E), showing a steady decline with depth. In contrast, Porto MV station 144 had exceptionally high numbers which were outside of the upper expected limit for subseafloor sediments at 0.9 mbsf (5.11 × 10⁹ cm⁻³ sediment) and then cell numbers declined rapidly and were close to the lower expected limit (Figure 2A). Interestingly, the low cell counts for most MV sediments below the SMTZ (0.2–0.5 mbsf) were equivalent to cell numbers found below 100 mbsf in other subsurface sediments (Supplementary Figure S1) based on the global subseafloor sediment trend (Parkes et al., 2000). This may suggest that prokaryotic populations from these depths are derived from populations at much greater depths and have been transported upwards during episodic MV eruptions. Such low numbers of cells presumably impacted on the downstream analysis of MV samples, as the majority of sediment samples analyzed for DNA analysis were not readily amplifiable by direct PCR with either archaeal or bacterial 16S rRNA gene primers (Supplementary Table S1).

Microbial populations from the methanogenic zones of different Gulf of Cádiz MVs were investigated by a combination of molecular methods including 16S rRNA and mcrA gene clone libraries, 16S rRNA gene tag sequencing and illumina 16S rRNA gene amplicon sequencing. However, DNA was only successfully amplified by direct PCR from seven different sampling stations covering four different MV sites (Darwin, Bonjardim, Captain Arutyunov and Porto MVs; Supplementary Figures S2–S4). Bacterial and archaeal 16S rRNA gene libraries made by different methods were generally in agreement with the exception of the Archaea V4-V5 region 16S rRNA gene amplicon library (Supplementary Figure S2) which differed considerably in identification of major archaeal taxonomic groups when compared with the other methods used. This may be due to PCR bias introduced by this set of archaeal primers (Teske and Sørensen, 2008).

Archaeal diversity at all four MV sites analyzed (Supplementary Figure S2) showed that Archaea involved in methane cycling (methanogens and ANME) were present at each site by at least one PCR-sequencing method. However, their proportions varied between MV sites and between stations of the same MV. For example, ANME-1 phylotypes at Captain Arutyunov MV were prevalent but varied with station; 88.4% at station 066, 52% at station 206, 39.2% at station 227 and 14% at station 191 when comparing the V2-V5 region 16S rRNA gene clone libraries. Whereas methanogens from the Euryarchaeota orders Methanosarcinales and Methanobacteriales were dominant at station 191 (65% of V2-V5 region 16S rRNA gene library) but were a minor component or absent at the other three Captain Arutyunov MV stations (0–6% of V2-V5 region 16S rRNA gene libraries). A high proportion (67.3%) of ANME-1 sequences were also present at station 191 in the V4-V5 region 16S rRNA gene library. Other major taxa at Captain Arutyunov included members of the Euryarchaeota belonging to Thermoprofundales (also known as Marine Benthic Group-D or Izemarchaea; Baker et al., 2020), Poseidonales (formerly Marine Group II; Rinke et al., 2019), Pontarchaea (formerly Marine Group III; Adam et al., 2017), Methanomicrobiales, Hadesarchaea (formerly SAGMEG; Baker et al., 2016) and Terrestrial Miscellaneous Euryarchaeotal group (Takai et al., 2001), along with members of the Bathyarchaeota (formerly Miscellaneous Crenarchaeotal Group; Meng et al., 2014), Lokiarchaeota (formerly Marine Benthic Group-B or Deep-Sea Archaeal Group; Baker et al., 2020), Parvarchaeota and Thaumarchaeota (including Marine Group 1 and Marine Benthic Group-A; Baker et al., 2020; Supplementary Figure S2).

Similar groups of Archaea were also present at Darwin MV station 038 and Bonjardim MV station 131; with Bonjardim MV having a high proportion of 16S rRNA gene sequences belonging to Methanosarcinales and Lokiarchaeota, and Darwin MV having a high number of ANME-1 and Bathyarchaeota 16S rRNA gene sequences, as well as Lokiarchaeota identified by V4-V5 16S rRNA gene amplicon sequencing. In contrast, Porto MV at 0.9 mbsf was dominated almost entirely by members of the ANME-1 with 96.8–100% of archaeal 16S rRNA gene sequences identified by all PCR-sequencing methods. Methanogens and ANME were also confirmed in all four MV sediments successfully analyzed for mcrA genes. ANME-1 and ANME-2 mcrA gene sequences were dominant in all MVs with the exception of Bonjardim MV which had 100% of mcrA gene sequences belonging to the methanogen genus Methanococcoides (Supplementary Figure S3).

Overall, bacterial diversity (Supplementary Figure S4) was higher than archaeal diversity with a greater average number of higher taxonomic groupings being identified per MV sediment analyzed (14 bacterial groups compared with 9 archaeal groups). However, all MVs were dominated or had a high proportion of Atribacterota (class JS1; Webster et al., 2004; Nobu et al., 2016), Chloroflexota, Pseudomonadota (formerly Proteobacteria), Deltaproteobacteria, Planctomycetota, Bacillota (formerly Firmicutes) and Ca. “Aminicenantes” (formerly OP8; Farag et al., 2014) with other bacterial lineages being less abundant (Supplementary Figure S4). Interestingly, Porto MV station 144 (0.9 mbsf) and Captain Arutyunov MV station 206 (0.7 mbsf) had up to 94 and 73% Atribacterota (class JS1) 16S rRNA gene sequences, respectively.

Methanogenesis from four different ¹⁴C-substrates (acetate, bicarbonate, methylamine and methanol) was detected at all MV sites (Carlos Ribeiro, Captain Arutyunov, Darwin and Mercator) analyzed, albeit at low rates (0.001–50.5 pmol cm⁻³ d⁻¹) and at discrete sediment depths (Supplementary Figure S5; Supplementary Table S5). However, despite these low rates, the average rate of methanogenesis for each MV was elevated when compared to the respective non-mud volcano reference site, with average MV methanogenic rates being up to two orders of magnitude higher (Supplementary Table S2). Maximum rates of methanogenesis tended to occur in the upper 0.5 m of MV sediments (Supplementary Figure S5) and at some sites this corresponded with decreases in sulphate concentrations (e.g., Carlos Ribeiro and Captain Arutyunov; Figure 2). Interestingly, the average proportion of combined methanogenic rates for the four MV sites due to hydrogenotrophic and acetoclastic methanogenesis (Supplementary Table S2) were similar at 22.8 and 23.7%, respectively, with methylotrophic rates using methylamine and methanol accounting for the remainder (48.9 and 4.6%, respectively). However, methylotrophic methanogenesis was less dominant in the reference station sediments, accounting for 35.9% with methylamine (and 0% from methanol) of the combined average methanogenic rates, with acetate and H₂/CO₂ being responsible for the remainder (33.3 and 30.8%, respectively; Supplementary Table S2).

Methanogenesis from added substrates (Table 2) was detected in 68 (28.3%) of the MV sediment slurries (total 240 sediment slurries) incubated for 130 days, with 50% of the methane producing slurry enrichments having been supplemented with the methylotrophic substrates, methanol and methylamine (15 and 19 enrichments, respectively). Acetoclastic methanogenesis represented 17.7% of the methanogenic slurries, while only 4.4% of the slurries produced methane with H₂/CO₂. The remaining methane producing slurries were supplemented with either benzoate (14.7%) or hexadecane (13.2%). An additional subset of sediment slurries (40 control sediment slurries) without substrates also showed methane production ranging from no detection to >1,000 nmol CH₄ (see Table 2), presumably due to methane degassing and/or methane production from in situ substrates, confirming the above methanogenesis activity measurements. This control methane value was then removed from the parallel amended slurries.

Closer inspection of the sediment slurries taking into account MV and sediment depth (Table 2) demonstrated that 55% of the different sediments tested (40 different sediments) had the potential to produce methane from methylamine or methanol. Whereas methane from other substrates was only found in 30 and 7.5% of sediments with acetate and H₂/CO₂, and 25 and 22.5% of sediments for benzoate and hexadecane, respectively. Interestingly, of all the different metabolic groups of methanogens tested for, only methylotrophic methanogenesis occurred at all seven MV sites across the Gulf of Cádiz. However, despite methylotrophic methanogenesis being prevalent in the Gulf of Cádiz MV sediments, methane production varied considerably depending on MV site and/or sediment depth. For example, methane produced ranged from 0.11–105.91 nmol CH₄ cm⁻³ d⁻¹ for methylamine and 0.08–118.24 nmol CH₄ cm⁻³ d⁻¹ for methanol, with highest methane production for both substrates being in sediments from Darwin MV at 0.11 mbsf (Figure 3; Table 2). Generally, rates of methylotrophic methanogenesis were highest in near surface sediments and decreased steadily with sediment depth at all MV sites (Figure 3B).

Similarly, potential methanogenesis also varied between MV sites and sediment depths for all other substrates tested, but methane production was often several orders of magnitude lower than with methylamine or methanol (Figure 3; Table 2). For example, methane produced by acetoclastic methanogenesis varied from 0.09–2.09 nmol CH₄ cm⁻³ d⁻¹, hydrogenotrophic methanogenesis ranged from 0.12–3.54 nmol CH₄ cm⁻³ d⁻¹ and hexadecane and benzoate were 0.1–4.51 and 0.12–3.16 nmol CH₄ cm⁻³ d⁻¹, respectively (Figure 3A; Table 2). Interestingly, the highest methane produced for these substrates occurred at different MVs to that with methylamine and methanol, with acetate, benzoate and hexadecane showing the highest rates of methanogenesis at Mercator MV, and H₂/CO₂ at Captain Arutyunov MV. Also, in contrast to methylotrophic methanogenesis, methane production for acetate, benzoate, hexadecane, and H₂/CO₂ did not show a clear trend with sediment depth (Figure 3A).

At the deep water (3,860 m) Porto MV, 58.3% of sediment slurry enrichments produced methane with 16.6% of these showing methane production higher than 100 nmol CH₄ after 130 days, whereas at the other two deep water (>2,000 m) sites (Bonjardim and Carlos Ribeiro MV), the numbers of positive methanogenic slurries were lower, with 20.8% at Bonjardim MV (12.5% with >100 nmol CH₄ after 130 days) and 16.6% at Carlos Ribeiro MV (0% with >100 nmol CH₄ after 130 days). The proportion of sediment slurries from the intermediate water depth (~1,000 m) MV sites, Captain Arutyunov MV and Darwin MV, producing methane were 26.7 and 16.6%, respectively, with a similar low percentage of these methanogenic slurries (16.7 and 11.1%, respectively) having methane production values >100 nmol CH₄ after 130 days. In contrast, the shallow Meknes MV site (694 m) showed both a high percentage of positive methane producing slurries (66.6%) and a high number (58.3%) of methanogenic slurries with methane production >100 nmol CH₄ after 130 days. However, Mercator MV, the shallowest MV site (346–470 m) which also had the largest number of sediments sampled (14 different sediments) for methanogenesis only had 28.6% positive methanogenic slurries, and 19% of these were >100 nmol CH₄ after 130 days. Such contrasting results suggest that the potential for MVs to produce methane is not linked with their seawater depth.

However, the sediment slurries that produced the highest amounts of methane (>100 nmol CH₄ after 130 days) from MVs with a water depth > 1,000 m (Porto, Bonjardim, Captain Arutyunov and Darwin MVs) were dominated by methylotrophic methanogenesis (15 out of 17 high-rate slurries; Table 2). Whereas MVs located below 1,000 m water depth (Meknes and Mercator MVs) were capable of producing high amounts of methane from a much broader range of metabolic substrates. For example, sediments from Meknes and Mercator MVs had only 7 out of 23 high-rate methanogenic slurries with added methylamine or methanol (Table 2).

Furthermore, increased incubation times (up to 300 days) led to several MV sediment slurries, which at 130 days showed no or low rates of methanogenesis (<10 nmol CH₄ after 130 days), developing the ability to generate high amounts of methane (>100 nmol CH₄ after 300 days). These were all from sediments of Carlos Ribeiro MV (0.55 and 1.05 mbsf) and Darwin MV (0.11 and 0.18 mbsf; Table 2), and the majority (83.3%) of these extended incubation time high-rate methanogenic slurries were incubated with the methanogenic substrates, methanol or methylamine.

Methane removal was also observed in a number of sediment slurries incubated with methanogenic substrates (Table 2) presumably due to anaerobic oxidation of methane (AOM). Interestingly, the number of slurries with detectable AOM (defined as >10 nmol CH₄ consumed after subtraction of unamended control slurry) were similar for each substrate tested (10–16 slurries per substrate) with the exception of H₂/CO₂ which accounted for 24.7% (21 slurries) of AOM slurries. The highest number of sediment slurries detected with AOM were obtained from sediments from Captain Arutyunov, Mercator and Carlos Ribeiro MVs with 27, 26 and 21 AOM slurries, respectively. Furthermore, Carlos Ribeiro MV sediments were particularly effective at removing methane as this MV had 70% (21 out of 30) of its sediment slurries showing AOM.

All MV sediment slurries showing relatively high rates of methanogenesis (>100 nmol CH₄) after 130 days incubation were screened for archaeal 16S rRNA and methanogen-specific mcrA genes using PCR-DGGE and sequencing (Figure 4). Despite a range of substrates shown to produce methane (Table 2), the majority of successful 16S rRNA gene PCR amplifications from extracted DNA were mainly obtained from sediments incubated with either methanol or methylamine. Presumably due to the potential rates of methanogenesis in these methylotrophic slurries often being 10-100-fold higher (Figure 3; Table 2) than those from other substrates, resulting in a stimulation of methanogen cell numbers and increased biomass and/or extractable DNA.

Sequencing excised archaeal 16S rRNA gene DGGE bands revealed that most of the methanogenic enrichments incubated with either methanol or methylamine were dominated by Archaea belonging to the methanogen order Methanosarcinales and were related (92–100% sequence similarity) to sequences previously found in a number of mud volcanoes and marine sediments (Supplementary Table S3), as well as pure cultures of Methanococcoides alaskense (97–99%) and Methanococcoides methylutens (99–100%). However, one methanogenic sediment slurry (Figure 4) from Porto MV (0.77 m) and incubated with methylamine did not contain detectable methanogens, but surprisingly was dominated by anaerobic methane-oxidizing Archaea ANME-1. These sequences were related (98% sequence similarity) to 16S rRNA gene sequences previously found in sulphate–methane transition zone (SMTZ) sediments from Aarhus Bay (Webster et al., 2011). Similar ANME-1 sequences were also found in Porto MV sediment enrichments incubated with benzoate (Figure 4), along with a second ANME-1 16S rRNA gene phylotype related to sequences from a mud volcano (Amsterdam MV) in the Eastern Mediterranean Sea (Pachiadaki et al., 2011). ANME-1 sequences were also the dominant archaeal group in the methanogenic slurry from Mercator MV near surface sediment (0.15 mbsf) incubated with hexadecane (Supplementary Table S3). However, it should be noted that the presence of ANME-1 within the enrichments could be due to their persistence in the sediment slurry rather than their use of the methanogen substrates directly.

Methanogen diversity in all methanogenic slurries were also investigated using the mcrA gene. The majority of methylotrophic sediment slurries analyzed by PCR-DGGE resulted in congruent mcrA gene sequences to those found by 16S rRNA gene analysis (Figure 4). Many slurries contained mcrA gene sequences that were 96–100% similar to those found in several Methanococcoides species and shared 97–99% sequence similarity to sequences previously found in marine sediments from Guaymas Basin, Marennes-Oléron Bay and Cascadia Margin (Dhillon et al., 2005; Roussel et al., 2009; Yoshioka et al., 2010). However, no mcrA gene sequences were found that corresponded with ANME-1 archaea, instead all ANME-related mcrA sequences identified in Gulf of Cádiz MV sediment slurries were related to sequences that grouped with ANME-2 (Supplementary Table S3). For example, Porto MV (0.77 mbsf) methanogenic sediment slurries incubated with benzoate or methylamine contained ANME-2a (mcrA group e) and/or ANME-2c (mcrA groupc/d) mcrA sequences similar (96–100% sequence similarity) to those found in sediments from Napoli MV (Lazar et al., 2011a) and Eel River Basin (Beal et al., 2009). Interestingly, methanogenic slurries from Mercator MV site 019 at depths 1.90 and 2.23 msf incubated with hexadecane contained mcrA gene sequences that were closely (100% sequence similarity) related to the hydrogenotrophic methanogen Methanobrevibacter arboriphilus.

In contrast to Archaea, the bacterial diversity within many of the methanogenic slurries was relatively high (Figure 4) with slurries containing sequences representative of several phyla (sometimes up to 5 phyla per slurry) and up to 11 detectable bacterial phylotypes (e.g., Bonjardim MV). Overall bacterial phylotypes belonged to 13 major taxa, including members of the Betaproteobacteria, Gammaproteobacteria, Deltaproteobacteria, Campylobacterota class Epsilonproteobacteria, Bacillota, Chloroflexota, Bacteroidota, Planctomycetota, Actinomycetota, Cyanobacteria, Atribacterota (JS1) and Ca. “Aminicenantes” (OP8), and one novel phylum level group previously found in deep marine sediments of the Japan Sea (Yanagawa et al., 2011). However, despite high bacterial diversity within the sediment slurries no clear pattern was observed to suggest that bacterial diversity was related to the methanogenic substrates added. Interestingly, sequences from Atribacterota, Ca. “Aminicenantes,” Actinomycetota, Planctomycetota and the novel bacterial group were only found in sediment slurries from MVs located in water depths >1,000 m, whereas Pseudomonadota and Bacillota were found in most MV sediments. It should also be noted that several methanogenic slurries dominated by Methanococcoides species contained no detectable Bacteria (e.g., Captain Arutyunov MV), and that many of the bacterial phylotypes that were detected in slurries were related to sequences previously found in other marine sediments (Supplementary Table S4), including those from cold seeps and submarine mud volcanoes (e.g., Heijs et al., 2008; Orcutt et al., 2010; Pachiadaki et al., 2011; Yanagawa et al., 2011).

MV sediment slurries (from Bonjardim, Carlos Ribeiro, Captain Arutyunov, Darwin, Meknes and Mercator) showing high methanogenic activity and detectable methanogens, by both 16S rRNA and mcrA genes, were used to isolate novel methanogens. After further enrichment, subculture and isolation (see Watkins et al., 2012, 2014) five out of six MV sediments yielded pure methanogen cultures. The majority of pure methanogens were able to utilise methylotrophic substrates and one was isolated on H₂/CO₂. All isolated methanogens belonged to the methylotrophic genus Methanococcoides (Table 3), with the exception of the one hydrogenotrophic Methanogenium species from the surface sediments of Darwin MV.

Our microbial diversity survey of mud volcanoes from the Gulf of Cádiz showed that sediments from below the SMTZ and within the methanogenic zone (below approx. 0.3 mbsf) have generally low numbers of prokaryotic cells, except Porto MV at 0.9 mbsf which had exceptionally high numbers (Supplementary Figures S6). Based on 16S rRNA gene sequences, microbial populations were dominated by methanogenic archaea (Methanosarcinales), anaerobic methane-oxidizing archaea (ANME), Bathyarchaeota, Lokiarchaeaota and bacteria such as, Atribacterota, Chloroflexota, Bacillota and Deltaproteobacteria (Supplementary Figures S2, S4, and S6) frequently found in other methane-rich sediments (Inagaki et al., 2006; Parkes et al., 2014; Vigneron et al., 2015; Katayama et al., 2016; Teske et al., 2021). Similar groups of archaea and bacteria have also been identified at submarine MVs from other locations including Håkon Mosby, Amsterdam, Napoli, Kazan, Chefren, Ryukyu Trench and Beaufort Sea MVs (Niemann et al., 2006b; Heijs et al., 2007; Omoregie et al., 2008; Lazar et al., 2011b, 2012; Pachiadaki and Kormas, 2013; Hoshino et al., 2017; Lee et al., 2019) and from the Gulf of Cádiz (Captain Arutyunov MV, Niemann et al., 2006a; Mercator MV, Maignien et al., 2013). For example, ANME-1 and ANME-2 were found to be the dominant archaea groups in sediments below the SMTZ (0.15 mbsf) at Amsterdam MV, ranging between 74 and 91% of the sequences found with only a few recognized methanogen groups (4% of sequences), whereas Deltaproteobacteria and Atribacterota accounted for 34 and 24%, respectively of the bacterial population (Pachiadaki et al., 2011). Similarly, ANME-1 and ANME-2 phylotypes dominated the sediments (0.4 mbsf) at Captain Arutyunov MV with members of the Deltaproteobacteria SEEP SRB-1 group (the AOM syntrophic partners of ANME-1 and ANME-2) accounting for 85% of bacterial sequences with a further 6% of sequences belonging to Atribacterota (Niemann et al., 2006a). Interestingly, in the study by Niemann et al. (2006a) a novel Bacillota clone CAMV300B902 was also identified, and similar 16S rRNA gene sequences were present in our study of Captain Arutyunov MV site 191 (0.45 mbsf) at high frequency (up to 55% in the V1-V5 16S rRNA gene library). Thus far, closely related sequences (>99% sequence similarity) to this phylotype have only been found at Captain Arutyunov and Kazan MVs (Niemann et al., 2006a; Pachiadaki et al., 2010), and their function is unknown. However, of note is that the phylotype’s nearest pure culture relative (90% 16S rRNA gene sequence similarity) is the SRB, Desulforudis audaxviator (Karnachuk et al., 2019), found in a deep South African gold mine, where it formed a near single-species ecosystem fueled by H₂ from water radiolysis (Chivian et al., 2008). This may suggest that novel Bacillota involved in hydrogen metabolism could be important bacteria at specific mud volcanoes. However, further investigations are necessary to determine the role of these bacteria.

Methanogens and methanotrophic archaea were also evident from analysis of the gene encoding for methyl coenzyme M reductase (mcrA), an enzyme which catalyzes the final step of the methanogenic pathway present in all methanogens and anaerobic methanotrophs (Hallam et al., 2003). The mcrA gene libraries were dominated by sequences belonging to ANME, with three (Darwin, Captain Arutyunov and Porto MV) out of the four mud volcanoes analyzed having 89–100% of mcrA gene sequences belonging to ANME-1 or ANME-2 lineages. Interestingly, only the gene library from Bonjardim MV was clearly dominated by methanogen mcrA genes, with 100% of sequences being related (93–98% sequence similarity) to mcrA genes found previously in gas hydrate sediments of the Cascadia Margin (Expedition IODP 1327; Yoshioka et al., 2010) and known cultured Methanococcoides species (Webster et al., 2019; Supplementary Figure S3). Similar Methanococcoides-related mcrA sequences were also present at Captain Arutyunov MV stations 206 and 191 but at much lower abundance, comprising 3–7% of the respective gene libraries. In support of the mcrA gene findings, low numbers of 16S rRNA gene sequences of Methanococcoides species were also identified in all Gulf of Cádiz MV sites analyzed here and ranged from 25% at Bonjardim MV to 0.01% at Porto MV (Supplementary Figure S2), and 1.2% in a separate study at Mercator MV (Maignien et al., 2013). In addition, other novel members of the Methanosarcinales were also identified that were related to 16S rRNA gene sequences from sediments of the Kazan MV (Heijs et al., 2007; Pachiadaki et al., 2010) and the terrestrial Lei-Gong-Huo MV (Chang et al., 2012).

Whilst measurements of methanogenesis in global marine sediments are relatively widespread there is comparatively little information from on or around mud volcanoes. Krüger et al. (2005) measured methanogenesis and methane oxidation at a wide variety of ocean sites, including the Håkon Mosby MV, using both in vitro and ¹⁴C-radiotracer techniques. This study found that methanogenic rates at Håkon Mosby MV were in the range of those found at other marine sediment sites, with 0.02–0.08 μmol CH₄ g⁻¹ sediment d⁻¹ (Krüger et al., 2005). However, these rates of methanogenesis were considerably higher than those found in our study of MV sediments of the Gulf of Cádiz (Supplementary Table S2; Supplementary Figure S5) and that by Maignien et al. (2013). Interestingly, an earlier study of the Håkon Mosby MV demonstrated that there was substantial variation (up to 4 orders of magnitude) in methanogenesis rates depending on where samples were located within the MV crater (Pimenov et al., 1999), with some rates being comparable to those measured at Gulf of Cádiz MVs. Clear differences in methanogenesis due to sample location were also observed in our study, demonstrating a high degree of variation between sample location at each MV site (Supplementary Figure S5). For example, Mercator MV had a peak rate of hydrogenotrophic methanogenesis of 50.54 pmol cm⁻³ d⁻¹ at station 013 located within the crater rim but no activity was detected at any other station, including sediment cores from the crater center (stations 009 and 019). Furthermore, low rates of methanogenesis were also reported for the submarine mud volcano KMV#5 in the Nankai accretionary complex where hydrogenotrophic and acetoclastic methanogenesis were 0.6–128 and 0.004–0.10 pmol cm⁻³ d⁻¹, respectively (Ijiri et al., 2018). But unlike KMV#5 where hydrogen was the dominant substrate for methanogenesis, it appears that methylotrophic methanogenesis is the dominant pathway with nearly 50% of methanogenesis across all MV sites at the Gulf of Cádiz using methylamine (Supplementary Table S2).

Despite the low rates of methanogenesis measured in this study we were able to enrich methanogens and measure potential rates of methanogenesis in vitro from a number of mud volcanoes and sediment depths using a range of different substrates. These sediment enrichments further confirm that MV sediments from the Gulf of Cádiz are able to carry out methanogenesis and that the majority of MVs are dominated by methylotrophic methanogenesis rather than hydrogenotrophic or acetoclastic methanogenesis.

All methane producing sediment slurries incubated with methylamine or methanol were dominated by Methanococcoides-related methanogens. These methanogens are obligate methylotrophs that utilise C1 compounds as methanogenic substrates (Liu and Whitman, 2008). Unlike acetoclastic and CO₂-reducing methanogens, methylotrophic methanogens are not outcompeted by SRB in sulphate-rich marine sediments for substrates (Purdy et al., 2003), enabling them to undertake methanogenesis in shallower anoxic sediments. In turn, this allows them to take advantage of methylated compounds (e.g., methanol and methylamines) that are typically released during the decay of organic osmolytes from marine organisms, including both prokaryotic and eukaryotic organisms present in marine sediments (Jones et al., 2019). In this study all positive methylotrophic sediment slurries were from the first 2 mbsf sediments, suggesting that in situ populations of Methanococcoides are abundant in shallow subsurface sediments of mud volcanoes in the Gulf of Cádiz, presumably where methylotrophic (non-competitive) substrates are available, and that methylotrophic methanogens are not outcompeted by SRB.

In addition, as sediment depths increased, potential in vitro rates of methylotrophic methanogenesis generally decreased, which may reflect decreasing numbers of viable Methanococcoides cells. For example, in coastal marine sediments it has been shown by quantitative real-time PCR that Methanococcoides 16S rRNA genes can represent up to 20% of the near surface (~10 cmbsf) archaeal population, and then decline to >1% in deeper sediments (Webster et al., 2015), similar to the Methanococcoides abundance values detected in 16S rRNA gene libraries in this study. Furthermore, Methanococcoides species were abundant in South China Sea sediments down to ~8.3 mbsf (Xu et al., 2021). Methanol and methylamines have been shown to be important substrates for methanogens in salt marsh sediments (Oremland et al., 1982), mangroves (Lyimo et al., 2009) and marine sediments (King et al., 1983; Zhuang et al., 2018; Xu et al., 2021), plus methylamines are the main methanogenic substrates in MV sediments from the Eastern Mediterranean Sea (Lazar et al., 2011b, 2012). Some Methanococcoides species isolated from the Gulf of Cádiz (Table 3) and the Napoli MV have also been shown to be capable of utilizing tri-, di and mono-methylamine and methanol, as well as other methylotrophic substrates like choline, dimethylethanolamine and glycine betaine directly (Watkins et al., 2012, 2014; L’Haridon et al., 2014). Studies of coastal salt marsh sediments have shown that closely related Methanococcoides species also undertake syntrophic partnerships with bacteria in a two-step process involving the formation of trimethylamine from glycine betaine (Jones et al., 2019) or choline (Jameson et al., 2019) to produce methane. It is conceivable that similar syntrophic relationships could also be occurring in MV sediments when more complex methylotrophic substrates are available, especially as glycine betaine-degrading bacteria are thought to be widespread in subsurface sediments (Jones et al., 2019). In the Gulf of Cádiz, sources of such compounds (methylamines, choline, and glycine betaine) could be made available to methanogens from the large populations of chemosynthetic macrofaunal communities that are often found around the crater of these MVs (Rodrigues et al., 2010; Cunha et al., 2013), as well as from the burial of microbial biofilms at the MV surface (Magalhães et al., 2012). Furthermore, detectable levels of methanol have also been found in sediments of several MVs including Captain Arutyunov and Carlos Ribeiro at depth (Supplementary Figure S7) as well as other subsurface sediments (Yanagawa et al., 2016; Zhuang et al., 2018; Xu et al., 2021), and methylamine has been shown to reversibly adsorb to clay minerals in sediments making it difficult to detect but remain biologically available to methylotrophic methanogens (Xiao et al., 2022). It maybe that such methanogens and methylotrophic methanogenesis is more widespread than previously thought (Valentine, 2011; Xu et al., 2021; Bueno de Mesquita et al., 2023) and that they are well suited to adapt to fluctuating geochemical environments, because of their ability to use a range of methylotrophic substrates. Furthermore, novel, and putative methylotrophic methanogens may also exist in the phyla Bathyarchaeota (Evans et al., 2015) and Verstraetearchaeota (Vanwonterghem et al., 2016), as well as the Euryarchaeota order Methanomassiliicoccales (Borrel et al., 2014), and may occur by ANME archaea (Bertram et al., 2013); all members of these archaeal groups can be identified in mud volcano sediments (Supplementary Figure S2). However, confirmation of methylotrophic methanogenesis within these groups can only be addressed by further investigations into the metagenomics or focused cultivation studies of MV sediments.

Hydrogenotrophic and acetoclastic methanogenesis activities occurred throughout sediments of the Gulf of Cádiz MVs, but potential rates were low (Supplementary Table S2). No methanogens were identified in any sediment slurries incubated with acetate or H₂/CO₂, and only six slurries (five with acetate) incubated with these two important methanogenic substrates produced methane that were > 100 nmol CH₄ after 130 days incubation (Table 2). This is in contrast to the generally accepted idea that methanogenesis in marine sediments is dominated by CO₂ reduction (Whiticar et al., 1986) and that acetate is an important substrate in deep subsurface sediments (Wellsbury et al., 2002). Hydrogenotrophic methanogenesis was reported to provide the majority of biogenic methane in the mud volcano KMV#5 from the Nankai accretionary complex, with active methanogenesis down to ~120 mbsf (Ijiri et al., 2018). Whereas, at the Amsterdam MV in the Eastern Mediterranean Sea, acetoclastic methanogenesis occurred, and was supported by high concentrations of acetate (up to 2 mM) in deeper sediment layers (Lazar et al., 2012). Similarly, methanogenic rates and mcrA gene sequences from an active brine seep mud volcano in the Gulf of Mexico revealed a predominance of acetoclastic over hydrogenotrophic methanogenesis (Joye et al., 2009). However, it maybe that longer incubations times (up to 2 years) are needed for batch enrichments of methanogens using these substrates (Katayama et al., 2022) or that more sophisticated methods similar to continuous-flow cultivation system used by Imachi et al. (2011) are necessary to culture methanogens from subseafloor sediments.

Methane production from hexadecane (Table 2), a long-chain alkane, in some methanogenic slurries (e.g., Mercator MV) demonstrated that methanogenesis from saturated hydrocarbons might be significant in MV sediments as previously shown in other anoxic environments (Zengler et al., 1999; Fowler et al., 2016). Methane from hexadecane often occurs by acetogenic bacteria decomposing alkanes to acetate and H₂, which in turn are available substrates for acetoclastic and hydrogenotrophic methanogens (Dolfing et al., 2008). In our study no direct evidence for such bacteria was observed in any hexadecane enrichment, although sequences related to acetogens (Acidaminobacter species; Stams and Hansen, 1984) were identified in other substrate-amended enrichments from the same MV sediment. However, syntrophic associations involving hydrogenotrophic methanogens and other bacteria, including SRB that can degrade hexadecane to CO₂ are also known (Brennan and Sanford, 2002). Intriguingly, Deltaproteobacteria 16S rRNA gene sequences related to SRB (Supplementary Table S4) and hydrogenotrophic Methanobrevibacter mcrA genes were found in hexadecane-amended Mercator MV enrichments (Supplementary Table S3), suggesting the presence of hexadecane-utilizing syntrophic associations. Turnover of hexadecane to methane was also shown to occur in hypersaline sediments of the Napoli MV (Lazar et al., 2011a,b), and terrestrial mud volcanoes from the Carpathian Mountains and Indonesia (Jiménez et al., 2016). It is possible that in high salinity MV sediments with elevated concentrations of alkanes, such as the Mercator MV (Perez-Garcia et al., 2011), methane production from alternative substrates is additionally important. Further experiments with increased incubation times (Katayama et al., 2022) may be necessary to determine the importance of alkanes as a substrate for methanogenesis in the Gulf of Cádiz.

Similarly, methane production from benzoate (Nottingham and Hungate, 1969) was also evident from sediments from several MVs, including Porto, Meknes and Mercator, but detection of the prokaryotes involved remained elusive. Only bacteria and archaea from the Porto MV sediment with benzoate were identified, with sequences mainly belonging to ANME and Atribacterota class JS1. Previous studies have shown that methanogenic benzoate degradation to carbon dioxide and methane is mediated by a consortium of bacteria (e.g., Syntrophus species) and hydrogen-utilizing methanogens (Stams and Plugge, 2009). It is therefore fascinating to think that a consortium of Atribacterota, thought to be heterotrophic (Nobu et al., 2016) and ANME which can produce methane by CO₂ reduction under certain conditions (Bertram et al., 2013) could be carrying out complete benzoate degradation through to methane. However, 16S rRNA gene sequences belonging to the genus Syntrophus (>0.01% of sequences) and hydrogenotrophic methanogens were found in sediments from all MVs analyzed from the Gulf of Cádiz, and previously 16S rRNA genes of Syntrophus species have been retrieved from the Kazan MV (Pachiadaki et al., 2010), suggesting that microbial communities in MVs have the potential to produce methane from aromatic compounds. The lack of a clear identification of methanogens from benzoate amended enrichments may also be due to the need for longer incubations or alternative methods of enrichment (Imachi et al., 2011; Katayama et al., 2022).