The surface water in the Levantine Basin of the SE Mediterranean Sea, including Israel's continental shelf, is an oligotrophic nutrient-poor marine system (Herut et al., 2000; Kress and Herut, 2001). The continental shelf narrows from south to north and is composed of Pliocene–Quaternary Nile-derived sediments. The sedimentation rate decreases with increasing distance from the Nile Delta and from the shoreline (Nir, 1984; Sandler and Herut, 2000). Off the shore of Israel, the sediment accumulation rate is relatively high at ∼ 0.1 cm yr-1 (Bareket et al., 2016). The bottom seawater along the continental shelf is well oxygenated and sulfate concentrations at the water–sediment interface are ∼ 30 mmol L-1 (Sela-Adler et al., 2015). The central and eastern regions of the Levantine Basin have relatively low total organic carbon (TOC) content (∼ 0.1 %–1.4 %; Almogi-Labin et al., 2009; Sela-Adler et al., 2015; Astrahan et al., 2017) as compared to the Western Mediterranean Basin and offshore the Nile River delta (1 %–2 %). Along the Egyptian coast, the TOC in surface sediments on the shelf reach maximum values of 1.5 % (Aly Salem et al., 2013). The finding of a free gas zone, which is located from a few meters to tens of meters below the seafloor (i.e., gas front), in seismic profiles within the sediments of the continental shelf of Israel (Schattner et al., 2012), led to the discovery of biogenic methane formation at some locations in the shallow sediments (Sela-Adler et al., 2015).

Seven sediment cores (∼ 5–6 m long) were collected using a Benthos 2175 piston corer, from the undisturbed sediments of the SE Mediterranean continental shelf of Israel at water depths of 81–89 m from three stations; SG-1, PC-3 and PC-5 (Fig. 1). The cores were sampled during cruises of the RV Shikmona between 2013 to 2017, and by the RV Bat-Galim on January 2017 (Table 1). The sediment cores were sliced on board every 25–35 cm within minutes upon retrieval from the seafloor. This area was previously investigated for other purposes, such as the sulfate reduction in the SMTZ (Antler et al., 2015; Wurgaft et al., 2019), and the possibility for methanogenesis (Sela-Adler et al., 2015).

From each interval, a 2.5 mL of sediment sample was collected and inserted immediately into an anoxic 10 mL glass bottle filled with 5 mL NaOH 1.5 N for headspace measurements of methane concentration (after Nüsslein et al., 2003). Approximately 3 mL of sediment was sampled every 50 cm for porosity. In addition, another 2.5 mL sediment sample was taken from each segment of the cores and transferred into a 20 mL glass bottle filled with a NaCl-saturated solution for H2 concentration measurements. Sediment samples from each segment of the cores were centrifuged on board if possible or in the lab within a day by Sorval centrifuge at 9299 g under 4 ∘C and Ar atmosphere in order to extract porewater for chemical analysis. The supernatant was filtered (0.22 µm) and analyzed for Fe(II), sulfate, sulfide, DIC and the stable carbon isotope composition of the DIC (δ13CDIC). After the porewater extraction, the sediment was analyzed for the content of the different reactive iron minerals. In addition, a sediment subsample from each segment of the January 2017 core from Station SG-1 was kept at -20 ∘C for molecular analysis. Due to high water content and movement in the uppermost part of the sediments, two ∼ 30 cm sediment cores were also subsampled separately, using a 0.0625 m2 box corer (Ocean Instruments BX 700 Al) and Perspex tubes, during the September 2015 and January 2017 cruises. The short cores were stored at 4 ∘C and cut in the lab within 24 h after their collection, and their results are presented for the top sediment (Fig. 2a–d).

The experimental setup (Table 2) consisted of 11 bottles with sediment from the methanic zone (265–285 cm depth) from Station SG-1, where iron reduction was apparent from the porewater profiles (Fig. 2d). Prior to the beginning of the experiment, sediment from the designated depth had been homogenized in an anoxic bag under N2 atmosphere. It was then transferred under anoxic conditions to a 250 mL glass bottle with the addition of synthetic seawater without sulfate to reach a 1 : 1 sediment : water slurry ratio for a 3-month incubation period. After the incubation period the slurry was subdivided anoxically to the 11 experiment bottles (60 mL each), and synthetic seawater was added for a final sediment : water ratio of 1 : 3. The bottles were sealed with a crimped cap and were flushed with N2 for 5 min, shaken vigorously and flushed again, (repeated 3 times). Three experimental bottles were autoclaved twice to serve as “killed” control for the experiment. The experimental bottles were amended with 1.6 g L-1 of hematite (Fe2O3) or 2.3 g L-1 of magnetite (Fe3O4) to reach Fe(III) final concentration of 10 mmol L-1. The three killed bottles were amended with the iron oxides after they cooled down to room temperature. H2 was added to some treatments to test its potential as an electron donor. A total of 1 mL of H2 was injected by a gas-tight syringe to the three killed bottles, to two bottles with the addition of hematite and to two bottles with the addition of magnetite (to reach a final concentration of ∼ 4 % of the head space volume). The experimental bottles were sampled several times for dissolved Fe(II) concentrations during the 14 d experiment period.

Geochemical porewater profiles of several sediment cores from the three stations (SG-1, PC-3 and PC-5; Fig. 1, Table 1) were produced in order to characterize the iron reduction process in the methanic zone of the SE Mediterranean continental shelf and to identify its potential sources. The porewater profiles at Station SG-1 (Fig. 2a) show complete depletion of total sulfur at approximately 150 cm depth in all cores. Sulfide concentrations were below the detection limit in all cores, indicating that the total sulfur is mostly sulfate. The methane concentrations in the porewater (Fig. 2b) show an increase with depth immediately after the consumption of sulfate. The maximum methane concentration was approximately 10 mmol L-1 at ∼ 140 cm depth in June 2015. The other methane depth profiles show an increase in the concentrations to approximately 2 mmol L-1 and then leveling off throughout the bottom of the cores (∼ 600 cm). Detected dissolved Fe(II) concentrations (Fig. 2d) were found in the upper iron reduction zone (between 30 and 90 cm depth), and a second peak was found in the deeper part of the sediment, at the methanic zone (below 180 cm depth). Maximum dissolved Fe(II) concentrations reached 84 µmol L-1 in the upper iron reduction zone of the sediments and 65 µmol L-1 in the methanic zone. The δ13CDIC values (Fig. 2c) were the lowest (-35 ‰), as expected at the SMTZ depth, and the highest in the methanic zone. H2 concentrations (Fig. 2e) decreased to a minimum of 0.017 µmol L-1 at 155 cm depth and then increased to a maximum of 0.147 µmol L-1 at 485 cm depth.

Porewater profiles from Station PC-3 (Fig. 2g–l) show similar patterns to Station SG-1 on all three sampling dates, but with lower methane concentrations. Total sulfur (Fig. 2g) was completely depleted within the upper 300 cm depth. Sulfide concentrations were below the detection limit at this station as well. Methane profiles show an increase in methane concentration immediately after the consumption of sulfate. The maximum methane concentration (Fig. 2h) reached 0.8 mmol L-1 at 450 cm depth in the core from August 2013. The dissolved Fe(II) profiles (Fig. 2j) show two peaks at this station as well, one in the upper part of the sediment with a maximum value of 32 µmol L-1 at 177 cm depth, and another one with a maximum value of 64 µmol L-1 at 390 cm depth at the methanic depth. The δ13CDIC values (Fig. 2i) decreased from approximately -10 ‰ at the water–sediment interface to -20 ‰ at the SMTZ. Below that zone there was an increase in δ13CDIC values to about -5 ‰ due to methanogenesis. H2 concentrations (Fig. 2k) remained around 2 µmol L-1 along the core. The three deviating points that do not fit the clear pattern are attributed to an analytical or sampling error.

Porewater profiles from the core collected at Station PC-5 (Fig. S1) resemble the profiles of Station PC-3. Total sulfur was depleted at approximately 300 cm, and methane concentrations increased below that depth to 0.3 mmol L-1. The Fe(II) profile shows two peaks in this core as well, one in the upper sediment of 20 µM at 150 cm depth and the second of 30 µmol L-1 in the methanic zone. The δ13CDIC value decreased from -5 ‰ at the water–sediment interface to -25 ‰ at the SMTZ, and below that depth increased to -17 ‰ in the methanic zone.

In addition to the dissolved constituents' profiles, reactive iron minerals were extracted from the sediment collected in 2015, and operationally defined iron mineral fraction profiles from Stations SG-1 and PC-3 were produced (Fig. 2f and l). In Station SG-1 there appears to be a slight variability in the content of the minerals (Fig. 2f). The Fecarb content in the upper part of the sediment was 0.22 dry wt %, increased to ∼ 0.45 dry wt % at 103 cm depth and then remained constant. The Feox1 content was 0.49 dry wt % in the upper part of the sediment, peaked at 203 cm depth to 0.64 dry wt % and then decreased to 0.50 dry wt % at the bottom of the core. The Feox2 content was 2.15 dry wt % in the upper part of the sediment, decreased to 1.03 dry wt % at 312 cm depth and then increased to 1.55 dry wt % at 427 cm depth. Femag content was 0.34 dry wt % in the upper part of the sediment, decreased to 0.32 dry wt % at 153 cm depth, increased to 0.35 at 253 cm depth, decreased to 0.23 dry wt % at 312 cm depth and increased again to 0.35 dry wt % at the bottom. A pyrite content profile from Station SG-1 was also produced (Wurgaft et al., 2019) from the September 2015 cruise and shows two peaks: the uppermost of 1.10 wt % at 153 cm depth, and the lower one of 1.80 wt % at 312 cm depth. The total reactive Fe(III) oxide profile showed a general decrease from 3.00 dry wt % at 13 cm depth to 2.27 dry wt % at 507 cm depth, with two minimum peaks of 2.42 dry wt % at 103 cm and of 1.88 dry wt % at 312 cm.

In Station PC-3 there appeared to be smaller changes in the different iron mineral fractions with depth (Fig. 2l). The Fecarb content in the upper part of the sediment was 0.50 dry wt % and reached 0.69 dry wt % in the deep sediment. The Feox1 content was approximately 1.00 dry wt % throughout the sediment column. The Feox2 content was 0.78 dry wt % in the upper part of the sediment, increased to 0.89 dry wt % at 167 cm depth and then decreased to 0.76 dry wt % at 495 cm depth. Femag content was 0.83 dry wt % in the upper part of the sediment, increased to 0.89 dry wt % at 167 cm and then decreased again to 0.75 dry wt % at 495 cm depth. The total reactive Fe(III) oxide content varied between 2.10 dry wt % (at 167 cm depth) and 1.76 dry wt % (at 137 cm depth).

Sediment from the observed deep iron reduction zone of Station SG-1 from the January 2017 core was used for a simple short-term (couple of weeks) slurry incubation experiment in order to characterize the iron reduction process in the methanic zone. Hematite and magnetite, which were expected to survive the sulfate zone, and were shown to be a source for AOM in lake sediments, were added to the slurries. Indeed, the operationally defined iron mineral fractions profiles (Fig. 2f) confirm that hematite and magnetite were abundant in the methanic zone in this core.

The results of the experiment are shown in Fig. 5. Dissolved Fe(II) concentrations show a significant increase from 11 µmol L-1 to approximately 90 µmol L-1 during the first 3 d in all the experimental bottles, except for the killed bottles, implying that the reduction is microbially mediated. Another observation was that the microorganisms were able to reduce both hematite and magnetite to the same extent. In addition, no difference in the Fe(II) concentrations between bottles with and without the addition of H2 was observed.

This study was performed in the SE Mediterranean (Fig. 1) above the area of a recently discovered gas front (Schattner et al., 2012). The investigated methane was found in the shallow sediments (∼ 1–5 m deep) and seems biogenic based on its low δ13CCH4 values and high C1 / C2 ratio (Sela-Adler et al., 2015). Station SG-1 is located at the center of the gas front area, while Stations PC-3 and PC-5 are located at the edges, and indeed methane related processes were more intensive at Station SG-1. The source of this gas front is not certain, but it was speculated to be terrestrial organic matter (Schattner et al., 2012). Our results suggest that there are two sources for methane in the shallow sediment: the first is from migration of methane from this gas front area (Wurgaft et al., 2019), and the second is from in situ methane formation, where the relative contribution of each source is currently unknown. In situ methanogenesis in the shallow shelf sediments is evident by the geochemical profiles of δ13CDIC and δ13CCH4 (Sela-Adler et al., 2015), by the microbial population abundance profile and by the functional mcrA gene profile (Figs. 3 and 4, further discussed below). The TOC content in the methanic zone is ∼ 0.8 % at Station SG-1 and ∼ 1 % at Station PC-3 (Sela-Adler et al., 2015), and these levels are known to be able to support in situ methanogenesis (Sivan et al., 2007).

The comparison between the sites show that methane reaches the highest concentrations at Station SG-1 (up to the saturation level; Sela-Adler et al., 2015), specifically in the June 2015 profile (Fig. 2b). This leads to intensive AOM by sulfate at the SMTZ, causing it to occur at shallower depth and to produce lower δ13CDIC values than the other two stations. The relation between the upward fluxes of methane, the SMT depth and the δ13CDIC values fit previous studies (e.g., Sivan et al., 2007). The higher methane concentrations in the June 2015 profile is presumably due to intensive migration of methane from the deeper sediments, and/or more intensive methane production at the exact location of the core collected at that time. The H2 concentrations at Station SG-1 (Fig. 2e) were lower by 2 orders of magnitude than the concentrations at Station PC-3 (Fig. 2k), perhaps due to more intensive hydrogen-consuming processes at Station SG-1 (i.e., sulfate reduction, methanogenesis, iron reduction; Conrad et al., 1986; Lovley, 1991). Dissolved Fe(II) porewater profiles (Fig. 2d and j) show some variability between the cores within the same station, probably as a result of environmental variations.

Despite the porewater profiles' variability between the stations, they show a resemblance in their trends. All geochemical porewater and iron mineral fraction profiles suggest that the sediments in this area of the SE Mediterranean shelf can be classified into three general depth zones (Fig. 2): zone 1 is the upper part of the sediment, where the classical iron reduction occurs, probably coupled to organic matter oxidation, with sulfate reduction below it; zone 2 is the SMT depth, where methane starts to increase, sulfate is completely depleted, and Fe(II) (Fig. 2d and j) is either present in low concentrations or absent (probably due to the precipitation of iron-sulfide minerals). In addition, the δ13CDIC values are the lowest in this zone, as expected from the intensive AOM process there, which uses the isotopically light carbon of the methane as a carbon source with small fractionation (Whiticar, 1999; Holler et al., 2009); zone 3 is the methanic zone, where methane concentrations increased to the highest values in all stations, as did the δ13CDIC since the carbon source for the methane comes mainly from CO2, leaving the residual DIC heavier by about 60 ‰ (Whiticar, 1999; Conrad, 2005). In this zone, local maxima of Fe(II) concentrations in the porewater were found in all cores, indicating reduction of iron oxides. The slurry experiment results show only a slight increase in Fe(II) concentrations in the killed bottles compared to their significant increase in the non-killed bottles, inferring that the iron reduction in zone 3 is microbial (Fig. 5).

This observed intensive iron reduction in the methanic sediments is the first discovered in the SE Mediterranean shelf. The phenomenon of iron reduction in the methanic depth has been observed before in other marine provinces (Jørgensen et al., 2004; März et al., 2008; Slomp et al., 2013; Riedinger et al., 2014; Treude et al., 2014; Oni et al., 2015; Egger et al., 2016). Yet, the type of link to the methane cycle is not well understood. Usually, iron reduction is coupled to oxidation of organic matter (Lovley and Phillips, 1988) and is performed by iron-reducing bacteria, which is probably the case in zone 1. It is, however, questionable whether this also stands for zone 3 and, if not, what process is responsible for the iron reduction at this depth and its relation to methane. The iron reduction in zone 3 can occur potentially via four pathways: (1) oxidation of organic matter arriving from the SMTZ, where it is produced by the microorganisms that live there and benefit from the upward migrating methane, (2) oxidation of the methane itself, (3) H2 oxidation or (4) oxidation of sulfur species through a cryptic cycle.

The oligotrophic nature of the water column in the studied area would suggest that intensive bacterial iron reduction coupled with the oxidation of organic matter in zone 3 is less likely. Nevertheless, we observe high methane concentrations in zone 3 in all three stations, where part of it is from upward migration. This indicates that regardless of the surface water oligotrophic nature, the TOC substrate may be enough to sustain all the microbial activity and to take part in the iron reduction process in the methanic zone. This is possibly due to biomass production in the SMTZ (i.e., the microbial community including anaerobic methanotrophs (ANMEs) and sulfate-reducing bacteria; Boetius et al., 2000) and its rapid use in the methanic zone (so the TOC content remains still low).

The importance of the methane flux as a carbon source that supports the deep microbial community in zone 2 and 3 in the sediments of the SE Mediterranean can be illustrated by comparing the organic carbon flux from the photic zone, with the flux of organic carbon that is oxidized by sulfate in the porewater. Using traps, Moutin and Raimbault (2002) estimated an export flux of 7.4±6.3 mg C m-2 d-1, which leaves the photic zone. However, Wurgaft et al. (2019) estimated that the flux of DIC toward the SMTZ from sulfate reduction is equivalent to 8±3 mg C m-2 d-1. Whereas the difference between the two fluxes is statistically insignificant, it should be noted that the flux of organic material that survives aerobic oxidation in the water column and the upper part of the sediment column, as well as anaerobic oxidation by other electron acceptors with a higher energy yield (Froelich et al., 1979; Emerson et al., 1980), is likely to be substantially smaller than the flux measured by Moutin and Raimbault (2002). Therefore, it is unlikely that export flux from the photic zone constitutes the sole source of carbon to the SMTZ. Wurgaft et al. (2019) suggested that “external” methane originates in deeper portions of the sediments and provides an important source of carbon to the SMTZ in Station SG-1. Such fluxes of “external” methane are common along continental margin sediments (e.g., Milkov and Sassen, 2002; Milkov, 2004; Zhang and Lanoil, 2004; Paull et al., 2008; Fischer et al., 2013). Here, we suggest that this supply of methane leads to intensive sulfate-mediated AOM in the SMTZ, and that this intensive process and biomass may serve as an additional substrate that “fuels” zone 3, activating the iron-oxides.

The recently discovered iron-coupled AOM process (Eq. 3) is the second potential process that can involve iron-oxide reduction in the deep methanic zone (Sivan et al., 2011: Segarra et al., 2013; Slomp et al., 2013; Riedinger et al., 2014; Egger et al., 2014, 2017; Rooze et al., 2016; Bar-Or et al., 2017). Fe(III) as an electron acceptor for AOM provides a greater free energy yield than sulfate (Zehnder and Brock, 1980), and its global importance was emphasized (Sivan et al., 2011, 2014; Segarra et al., 2013). Two of the main environmental conditions for iron-coupled AOM to occur are high dissolved methane concentrations and abundant reducible iron oxides (Riedinger et al., 2005, 2014; Egger et al., 2017). Thus, from our profiles it seems that AOM could be a valid option, considering the high methane concentrations and the high sedimentation rates (0.1 cm yr-1; Bareket et al., 2016), which allow the iron oxides to survive the sulfidic zone and reach the methanic zone (Riedinger et al., 2005, 2014; Egger et al., 2017). This can also be inferred from Fig. 6, where some association was observed between the dissolved Fe(II) concentrations and the methane concentrations in zone 3. It seems that at high concentrations of Fe(II), methane concentrations are low and vice versa. This could be a result of iron-coupled AOM that uses methane to reduce Fe(III)-oxides, releasing dissolved Fe(II) to the porewater. It can also suggest a type of competitive relationship between methanogenesis and microbial iron reduction, or a microbial population switching from methanogenesis to iron reduction metabolism (e.g., Sivan et al., 2016). It should be noted that our experiment was not designed to test AOM due to its short timescale of a few weeks, and hence another long experiment with the addition of the 13C-labeled methane will be needed to shed more light on this association.

The third potential process that can be coupled to iron reduction in the methanic zone is H2 oxidation. H2 is an important intermediate in anoxic aquatic sediments. In this type of environment, it is produced mainly by fermentation of organic matter (Chen et al., 2006) and can be involved in different microbial processes where each process would need a certain amount of H2 in order to occur (Lovley and Goodwin, 1988). The H2 levels at Stations SG-1 and PC-3 (Fig. 2e and k) are relatively high in comparison to other marine environments (Lilley et al., 1982; Novelli et al., 1987), suggesting that there is enough H2 to sustain the iron reduction process. The relatively high H2 concentrations at these stations could be explained by the dominance of H2 production processes (i.e., fermentation; Chen et al., 2006) compared to H2-consuming processes (i.e., sulfate reduction, methanogenesis, iron reduction; Conrad et al., 1986; Lovley, 1991). At Station PC-3, the H2 concentrations (Fig. 2k) are constant in zone 3, this suggest that in addition to being produced, H2 is consumed as well. At Station SG-1 (Fig. 2e) there is a maximum peak in zone 3, indicating that there is either more H2 production or less H2 consumption in this zone compared to zone 2. This is reasonable considering the intensive microbial activity in zone 2. The decrease in the H2 concentrations below the peak suggests that H2 consuming processes are intensive in this zone. The H2 involvement was tested by injecting 1 mL of this gas into the experimental bottles in the methanic iron reduction process (Fig. 5). We observed that the increase in the Fe(II) concentration was similar in the bottles with a H2 addition compared to the bottles without H2. This could mean that either there is enough H2 in the sediments as it is, as implied by the H2 porewater profiles, or that at the methanic depth H2 is not involved in the iron reduction process.

The fourth potential way to reduce iron in zone 3 is by an active sulfur cycle. The pyrite profile supports this possibility by showing two peaks, the uppermost in zone 2 of ∼ 1 wt % and the other in zone 3 of ∼ 2 wt % at about 300 cm depth (Fig. 2f). The peak at 300 cm depth indicates a possible active sulfur cycle, even though sulfate is already undetected at 200 cm. Thus, a possible scenario is that Fe(III) is reduced by pyrite oxidation (Eq. 3) (Bottrell et al., 2000), which triggers the “cryptic” sulfur cycle, as observed in other marine sediments (Holmkvist et al., 2011; Brunner et al., 2016; Egger et al., 2016). In this cycle, elemental sulfur, and eventually by disproportionation also sulfide and sulfate, is produced. The sulfide reacts with iron-oxide and precipitates as FeS or as pyrite (FeS2) (Holmkvist et al., 2011). The sulfate can inhibit methanogenesis (Mountfort et al., 1980; Mountfort and Asher, 1981), which can result in the enhancement of the iron reduction process due to competition for the substrate with the methanogenesis process. Another indication for an active sulfur cryptic cycle comes from the 16S rRNA sequencing analysis (Fig. 4), which shows that Proteobacteria, a potential sulfur-related bacteria phylum, is one of the most abundant phyla in the sediments. Moreover, the increase in the abundance of Sva0485 order of the Deltaproteobacteria class, a known sulfate reducer (Tan et al., 2019), with depth supports an active sulfur cycle in zone 3 as well.

Our data profiles and incubations indicate that the observed iron reduction in the methanic zone of the SE Mediterranean shelf is performed by microbial activity. The microbial results show first that the abundances of the bacteria and archaea (Fig. 4) are typical to oligotrophic marine sediments (e.g., South China Sea that contains ∼ 0.5 %–1 % TOC, Yu et al., 2018). Second, even though potential bacterial iron reducers, such as Alicyclobacillus, Sulfobacillusin, Desulfotomaculum genera (Firmicutes), Acidiphilium (Alphaproteobacteria), Desulfobulbus, Desulfuromonas, Geobacter, Geothermobacter, Anaeromyxobacter (Deltaproteobacteria) and Shewanella (Gammaproteobacteria) (Weber et al., 2006) comprise less than 0.1 % of bacteria detected in the methanic zone (from 185 cm and below), it appears that both the microbial abundance and the Fe(II) concentration peaked in this zone. Cultivation efforts indicated that archaeal methanogens may also play a role in iron reduction within sediments (Sivan et al., 2016). Moreover, the relative abundance of methane-metabolizing archaea was shown to correlate with Fe(II) concentrations in Helgoland muds from the North Sea, where microbial abundance and the Fe(II) concentrations peaked in the methanic zone (Oni et al., 2015), similarly to the results found in the SE Mediterranean sediments. It is possible that methane-metabolizing archaea were involved in the iron reduction in the SE Mediterranean sediments, as the highest mcrA gene copies per gram of wet sediment were detected in the SMTZ and at the top of the methanic zone (Fig. 3) where the Fe(II) concentrations are high (Fig. 2d). Methanotrophs, such as ANMEs, were found to be involved in iron-coupled AOM in marine and freshwater cultures (Scheller et al., 2016; McGlynn et al., 2015; Ettwig et al., 2016; Cai et al., 2018). ANMEs were found here with relatively low frequencies (ANME1, below 1 % in most samples, circa 5 % in the 185 cm layer), and their role in iron reduction within the SE Mediterranean sediments remains to be tested.

In our study, Spearman correlation analysis at Station SG-1 (Table S2) revealed that bacterial phyla SBR1093 (candidate phylum), Acidobacteria and Chloroflexi, as well as archaeal phylum Bathyarchaeota, showed significant positive correlation with a Fe(II) concentration in the methanic zone. The candidate phylum SBR1093 was first identified in phosphate-removing activated sludge from a sequencing batch reactor (Bond et al., 1995) and is often detected in a short-chain fatty-acid-rich environment such as wastewater treatment, and marine sediments (Wang et al., 2014). It was thought to be capable of growing autotrophically, but the metabolic capabilities related to iron reduction remain unclear. Strains of Acidobacteria and Chloroflexi phyla were found to be capable of iron reduction (Kawaichi et al., 2013; Kulichevskaya et al., 2014). In addition, members of Acidobacteria were found in iron-coupled AOM enrichment (Beal et al., 2009). The metabolic properties of subgroup 26 from Acidobacteria and TK10 Class of Chloroflexi are still not known. Bathyarchaeota are globally distributed and account for a considerable fraction of the archaeal communities in the marine sediments, particularly in the Mediterranean Pleistocene sapropels (Coolen et al., 2002; Zhou et al., 2018). While Bathyarchaeota have diverse metabolic capabilities (Lloyd et al., 2013; Meng et al., 2014; Evans et al., 2015; He et al., 2016; Yu et al., 2018; Feng et al., 2019), their role in iron reduction warrants further studies, as suggested from their high abundance here. Therefore, iron reduction and methane cycling within the deep methanic zone may be facilitated by an interplay among bacterial and archaeal groups, whose physiology and functions need further investigation.