Characteristics and Evolution of sill-driven off-axis hydrothermalism in Guaymas Basin – the Ringvent site

The Guaymas Basin spreading center, at 2000 m depth in the Gulf of California, is overlain by a thick sedimentary cover. Across the basin, localized temperature anomalies, with active methane venting and seep fauna exist in response to magma emplacement into sediments. These sites evolve over thousands of years as magma freezes into doleritic sills and the system cools. Although several cool sites resembling cold seeps have been characterized, the hydrothermally active stage of an off-axis site was lacking good examples. Here, we present a multidisciplinary characterization of Ringvent, an ~1 km wide circular mound where hydrothermal activity persists ~28 km northwest of the spreading center. Ringvent provides a new type of intermediate-stage hydrothermal system where off-axis hydrothermal activity has attenuated since its formation, but remains evident in thermal anomalies, hydrothermal biota coexisting with seep fauna, and porewater biogeochemical signatures indicative of hydrothermal circulation. Due to their broad potential distribution, small size and limited life span, such sites are hard to find and characterize, but they provide critical missing links to understand the complex evolution of hydrothermal systems.


Results
Mapping Ringvent. Ringvent is located 28.5 km northwest of the northern Guaymas Basin spreading center, in the central Gulf of California ( Supplementary Fig. 1). A seismic transect across the site shows a sedimented basin surrounded by a ring-shaped mound (Fig. 1). The bathymetric map shows that the ring-shaped mound rises to ca. 15 to 20 meters above the surrounding seafloor, with greater topographical highs in the western portion. Subbottom seismic imaging of the Ringvent area shows shallow reflective sedimentary layers on the outside of the ring structure and within the sedimented basin; specific reflectors intensify and approach the sediment surface near the topographical high of the ring structure. Underneath the ring, seismic energy is absorbed by free gas present in seismically blanked dome-shaped intrusions; the tops of these gas intrusions underlie the crests of the ring. The extent of shallow gas and the up-dip brightening of reflectivity towards the blanked zones deeper in the subsurface suggest ongoing gas and fluid transport (Fig. 1).
Bathymetric mapping by AUV Sentry shows large 5-to 10-m deep gullies incised laterally into the ring crest, and similar sedimented hollows on top of the crest ( Supplementary Fig. 2). These locations are characterized by elevated heat flow, water column redox and thermal anomalies ( Supplementary Fig. 3), exposed authigenic mineral concretions, widespread Lamellibrachia tubeworm colonies, localized venting of warm fluids, and mats of chemosynthetic sulfur-oxidizing bacteria and individual sulfur-dependent Riftia tubeworms that resemble their counterparts in the Guaymas Basin spreading center (Fig. 2). Photomosaic surveys by Sentry of the western portion of the ring-shaped mound reveal small outcrops and rocks overgrown with yellow and white mats of sulfur-oxidizing bacteria 14 , and cracks in the hard substrate that are populated with galatheid crabs, and tubeworms resembling Lamellibrachia ( Supplementary Fig. 4). Thin sediments of a few centimeters on the ring structure accumulate locally in sedimented hollows and incised gullies, and allow coring and heatflow measurements (Mound 1 and 2 site, ORP site; Supplementary Fig. 2). Thermal gradients in these surficial sediments measured with Alvin's heat flow probe reach approximately 5 °C/m, exceeding thermal gradients in seafloor sediments at the outer perimeter of the ring (0.6-0.7 °C/m) by an order of magnitude (Table 1, and Supplementary Fig. 5). The moderately warm surficial sediments in sedimented spots on the Ringvent mound contain porewater sulfide at concentrations of several hundred micromolar (Supplemental Data 1). Localized hydrothermal fluid flow with significantly elevated temperatures (ca. 20 °C to 70 °C, Tables 1 and 2) coincides with white and yellow-orange mats of large, sulfur-oxidizing filamentous bacteria of the family Beggiatoaceae -a microbial indicator of sulfide-rich, reducing hydrothermal flow 14 -and individual Riftia tubeworms surrounded by shimmering warm water of ca. 20 °C (Fig. 2). Riftia require turbulent mixing of sulfidic vent fluids with oxic seawater to sustain their sulfur-oxidizing, CO 2 -fixing bacterial endobionts 15 ; in Guaymas Basin they may also be able to use thiosulfate emerging from the sediments as a sulfur source 16 . Silica mineral concretions. Exposed seafloor mineral deposits were sampled at the ORP and Mound 1 locations for petrographic and compositional analysis ( Supplementary Fig. 6). Scanning electron microscopy/ energy dispersive X-ray spectroscopy (SEM/EDS) and X-ray diffraction analysis show that these deposits consist predominantly of a silica matrix formed by opal-A microspheres embedding diatom frustules, with subordinate amounts of low magnesium calcite, fibrous aggregates of aragonite, and platy single crystals and rosette aggregates of barite (Fig. 3). These minerals strongly differ from the classical cold seep carbonate deposits, formed through the precipitation of carbonates by microbially-mediated methane oxidation. The dominance of silica in the Ringvent minerals was confirmed by SEM/EDS analyses of Ringvent samples at the Woods Hole Oceanographic Institution ( Supplementary Fig. 7).
Silica precipitates were also collected by piston coring from the subsurface sediments in the central lower-lying Ringvent area ( Supplementary Fig. 8). A silica nodule from 2.35 m depth, recovered at this location by piston coring, has a bead-like texture of non-crystalline opal-A microspheres (Fig. 3G) that resembles laminated silica-carbonate hydrothermal deposits from intertidal hot springs with temperatures of 54 to 87 °C, at the eastern coast of the Baja California peninsula 17 . The petrographic identification of the opal-A silica was confirmed by XRD analysis (Supplementary Data 2). Very similar laminated structures -microbial filaments coated with amorphous silica -occurring in Guaymas Basin siliceous deposits are interpreted as the result of conductive cooling, without silica dilution by inmixing of seawater 18 . These similarities suggest that the silica-rich deposits of Ringvent form by diffuse flow and conductive cooling of silica-rich hydrothermal solutions.
Carbonate composition and isotopic signatures. Since hydrothermal fluids in Guaymas Basin are rich in methane and dissolved inorganic carbon (DIC) 19 , and methane-derived carbonates are widespread in the northern Gulf of California 20,21 , we checked piston-cored subsurface sediments for indicators of methane cycling and a potential methane contribution to carbonate formation. In core P11 from the central Ringvent area, conspicuous non-skeletal (i.e., not derived from plankton shells) carbonate concretions are found below ca. 50 cm depth, forming a massive layer at ca. 1 to 1.20 m below the seafloor ( Supplementary Fig. 8). This layer is unique to core P11; it does not appear in core P10, collected ca. 1 mile further west outside Ringvent, and also not in any other core collected during the 2014 Expedition (details in Methods: Field surveys). XRD examination (Supplemental Data 2) shows that these carbonates consist of magnesian calcite, compositionally similar to cold seep carbonates reported previously for Guaymas Basin 22 . These localized carbonate concretions exhibit light δ 13 C values from −37 to −41‰, whereas smaller, visually inconspicuous carbonate grains and particles in sediment surrounding these nodules show slightly heavier δ 13 C values from −37 to −30‰ (Figs 4 and 5A, and Supplementary Data 3). Carbonate nodules and grains in core P11 are significantly lighter than organic carbon in the same core, between −20.5 and −22‰ (see Methods), and previously analyzed sedimentary organic carbon and chemosynthetic microbial biomass (−21.6 and −22.1‰, respectively) in Guaymas Basin 23 , indicative of       Fig. 4), were not found in the sedimentary carbonates and nodules of core P11, but in fibrous aragonite veins embedded in Ringvent silicate samples ( Fig. 3) that were collected at a biologically active location at the Mound 1 site ( Supplementary Fig. 6). These findings indicate that anaerobic methane oxidation occurs locally in the silica-dominated matrix of Ringvent, and imply the presence of methane and methane-oxidizing microorganisms.
Carbonate origins. Potential formation conditions that inform the origins of the solid-phase carbonates in the Ringvent sediments were further constrained by oxygen isotope measurements. In the δ 13 C vs. δ 18 O plot of the carbonate nodules and sedimentary carbonates (Fig. 4), the most 13 C-depleted carbonates (fibrous aragonite; nodules and grains; clam shells) showed a moderate enrichment in 18 20 , calculated using the experimental equation for this mineral 24 . Notably, δ 13 C and δ 18 O values co-vary strongly in the carbonates from core P11 (R: −0.8) indicating a mixing line between two end-members: methane-derived carbonates precipitated in isotopic equilibrium with relatively cold bottom waters, and biogenic carbonates in warmer surface seawater. Indeed, the petrographic analysis of the sediments confirms that biogenic carbonate (coccoliths, benthic and planktonic foraminifera) is the dominant carbonate phase in sediment outside the nodule-bearing interval. Carbonates collected in core P11 outside the nodule-bearing interval, and from cores outside Ringvent, displayed limited variation of δ 18 O, between −1.1 and 1.8‰ ( Fig. 4; Supplementary Data 3), similar to typical values of biogenic carbonates in this region 25,26 .
To date the buried silica and carbonate deposits, we determined the local sedimentation rates based on the 14 C-age of sedimentary organic matter, undistorted by methane or DIC-derived contributions (see Methods), for the piston coring locations of the 2014 cruise (R/V El Puma, Supplementary Data 4). The local rate of 0.286 mm/y for core P11 was the slowest sedimentation rate obtained for all El Puma cores, and slower than previously published rates of 0.88 to 1.2 mm/yr for the Baja California slopes of Guaymas Basin 27,28 . This low sedimentation rate may result from reduced terrestrial contributions at this central Guaymas Basin location, but input of fossil organic matter into sedimentary organic matter may also play a role. Independently measured sedimentation rates for several Guaymas Basin locations also showed the lowest rates (0.5 mm/yr) at Ringvent 4 .
Porewater methane and DIC. Since the methane-rich hydrothermal and seep fluids of Guaymas Basin and the Sonora Margin carry the carbon isotopic imprints of different biological and hydrothermal sources, we examined porewater methane at Ringvent and other off-axis locations in Guaymas Basin. Porewater methane concentrations in core P11 range from near 1 to 1.5 mM, and gradually decreased to background levels within the upper 1.5 to 2 m (Supplementary Data 5).
Although decreasing methane concentrations suggest oxidation, the δ 13 C-CH 4 profile in the core P11 does not indicate a specific, localized sediment horizon where methane oxidation generates heavier δ 13 C-CH 4 signatures, as seen in the Sonora Margin cores at ca. 1.5 to 2 m depth (Fig. 5B), and possibly in an independently collected short pushcore from Ringvent (termed RingSeep) 4 , where a small methane pool (2 to 36 µM) changes from −57‰ at 20 cm depth towards −44.8‰ at the surface 4 . Instead, the δ 13 C-CH 4 values remained consistently within a narrow range, from −60 to −66‰ (Fig. 5B), between microbially produced methane (−97.3 to −82‰) in fully reduced, cold subsurface sediments on the nearby Sonora Margin 29 , and in Sonora Margin cores P06 and P12 (−89.1 to −75.8‰), and methane in hot hydrothermal surface sediments (near −42‰) 19 Supplementary Fig. 5). The temperature probe could not be fully inserted into the mineral mound, and the top sensor remained ca. 5 cm above the ground. (2019) 9:13847 | https://doi.org/10.1038/s41598-019-50200-5 www.nature.com/scientificreports www.nature.com/scientificreports/ closest matches to Ringvent are the intermediate δ 13 C-CH 4 values near −60‰ in deep subsurface sediments of Guaymas Basin (DSDP hole 481) that are interpreted as mixtures of hydrothermal methane originating at deep sills and microbially produced methane in the upper sediment column 31,32 , and δ 13 C-CH 4 values clustering near −55‰ in shallow sediments at several off-axis Guaymas Basin seep locations 4 . Near the top of core P11, residual methane is more strongly depleted in 13 C, with values near −66‰, indicating the influence of methylotrophic methanogenesis that competes successfully with sulfate reduction in surficial marine sediments 29 . Methylotrophic methanogenesis is also compatible with the isotopic composition of trace methane (5 to 14 µM, Supplementary Data 5) observed throughout sediment core P10, collected outside Ringvent (Fig. 5B).
Although methane oxidation would generate 13 C-depleted DIC, the δ 13 C-DIC values for core P11 stand out as the heaviest of all cores (Fig. 5C), and overlap with the δ 13 C-DIC range (−6 to +2.7‰) of seawater-impacted hydrothermal fluids 33 and porewater from shallow hydrothermal sediments in Guaymas Basin 19 . Instead, these results suggest mixing of seawater DIC (−0.6‰) and hydrothermally derived DIC (−9.4‰) 22 . Seawater influence is also reflected by the low porewater DIC concentrations that increase from 2.5 mM, resembling Guaymas bottom water (2.34 mM) 23 , towards 4 mM downcore (Supplementary Data 5). Seawater inmixing and ventilation resolve the apparent contradiction that decreasing methane concentrations towards the surface are not accompanied by carbon-isotopic signatures of porewater methane oxidation.
Sulfur cycling. Of all sediment cores in this study, core P11 has the highest porewater sulfate concentrations (28-29 mM), similar to seawater, throughout the entire depth range (Fig. 5D). Sulfide porewater concentrations are consistently below the detection limit (ca.  www.nature.com/scientificreports www.nature.com/scientificreports/ Margin cores P06 and P12 harbor very high biogenic sulfide concentrations reaching 5 to 10 mM (Supplementary Data 5). The isotopic composition of porewater sulfate was analyzed to obtain additional evidence for microbial sulfur cycling. Microbial sulfate reduction to sulfide discriminates against the heavy sulfur isotope 34 S, enriching residual sulfate in 34 S; the effect is reversed by oxidative sulfur cycling which returns isotopically light sulfur to the sulfate pool 34 . In contrast to the surficial sediments in Sonora Margin cores P06 and P12, which show massive biogenic 34 S enrichment (δ 34 S-SO 4 2− towards ≥80‰), the δ 34 S-SO 4 2− profiles of cores P10 and P11 change only minimally and thus indicate limited sulfate-reducing activity (Fig. 5E, and Supplementary Data 5). In core P10, the surficial sediments show moderate 34 S-enrichment and a very gradual increase towards δ 34 S of ~25.5‰ at depth, indicating that sulfate-reducing activity is limited to surficial sediments, likely fueled by planktonic organic matter. In contrast, core P11 displays an almost linear transition from isotopic compositions close to δ 34 S values of seawater sulfate (21‰) at the sediment surface towards δ 34 S values near 23‰ at the bottom of the core (Supplementary Data 5), suggesting mixing of surficial seawater and a subsurface sulfate pool. Together, the biogeochemical indicators of sulfur cycling in the central Ringvent sediments indicate a current regime of seawater inmixing, low sulfate-reducing activity, and lack of sulfide accumulation, which contrasts with the localized sulfidic conditions on the Ringvent mound that sustain sulfide-oxidizing microbial mats and symbiont-dependent invertebrates. The lack of sulfide likely inhibits anaerobic methane-oxidizing archaea, which generally require reducing conditions to be active 35 , and is consistent with methane removal predominantly by seawater inmixing and dilution.

Microbial communities.
To identify potential microbial catalysts of methane and sulfur cycling, and organic matter remineralization, the bacterial and archaeal communities in sediment samples from cores P10 and P11 were analyzed by high-throughput 16S rRNA gene sequencing (Fig. 6). Lineages of anaerobic heterotrophs dominate these communities; specifically, the bacterial phyla Chloroflexi (Anaerolineae and Dehalococcoidia) and Atribacteria (OP9/JS1 lineage), and the archaeal phyla Bathyarchaeota and Lokiarchaeota (Fig. 6A,B). Reconstructed genomes of these largely uncultured bacterial and archaeal lineages reveal the diverse, largely heterotrophic metabolisms of these benthic anaerobes 36,37 . Rarefaction analyses show that bacterial and archaeal community richness in the P11 samples are lower relative to all other piston-cored Guaymas sediment sites (Fig. 6C,D).
Archaea and Bacteria central to methane and sulfur cycling in the Ringvent data include methane-oxidizing ANME-1 archaea, and sulfate-reducing Deltaproteobacteria, in particular the family Desulfobacteraceae, the Desulfatiglans lineage, and an uncultured lineage that taxonomy servers assign to the family Desulfarculaceae (Supplementary Figs 9 and 10; Supplementary Data 6). While these methane-and sulfur-cycling anaerobes dominate sequencing surveys in hot hydrothermal sediments of Guaymas Basin 35,38 , they occur here in lower proportions. ANME-1 archaea and deltaproteobacterial sulfate reducers each account for approximately 4% of the sequence dataset in P11, compared to traces (ca. 0.01%) for ANME archaea, and ca. 3% for deltaproteobacterial sulfate reducers in core P10 (Supplementary Data 6). The sulfidic, methane-rich cores P6 and P12 yielded higher percentages of ANMEs, and reduced proportions (0.75 to 1.5%) of deltaproteobacterial sulfate reducers, possibly as a result of sampling below the sulfate-rich zone. The Desulfobacteraceae are a physiologically www.nature.com/scientificreports www.nature.com/scientificreports/ www.nature.com/scientificreports www.nature.com/scientificreports/ diversified family of sulfate reducers that oxidize low-molecular weight organic acids, acetate, and some aromatic compounds to CO 2 and occur frequently in marine sediments 39 . In core P11, these sequences include -in small numbers -members of the SEEP-SRB1a cluster 40 , the bacterial syntrophs of ANME methane oxidizers in marine sediments (Supplementary Fig. 10). The Desulfatiglans lineage, named for the cultured genus and species Desulfatiglans anilinii, includes aromatic-oxidizing isolates and enrichment cultures, and is widespread in marine sediments; subsurface-adapted metabolic capabilities, such as dehalogenation, also characterize this group 41 . The facultatively autotrophic sulfate-reducing family Desulfarculaceae 42 turned out to be unrelated to the Guaymas Basin amplicons that were assigned to this family by web-based taxonomy pipelines; instead, the amplicons formed a distinct phylogenetic lineage with clones from geographically diverse marine subsurface sediments ( Supplementary Fig. 10). Interestingly, cultured methanogenic lineages 43 , such as the hydrogenotrophic, autotrophic Methanocellaceae, the acetoclastic/methylotrophic Methanosarcinaceae, and the methylotrophic Methermicoccaceae, account only for small numbers of sequences in all piston-cored Guaymas sediments (Supplementary Data 6), and are unlikely to represent major methane producers in these sediments. Sediment pushcores that were collected by HOV Alvin from hydrothermally active sediments at the Ringvent structure (Mound 1 site) and analyzed specifically for methyl coenzyme M reductase (alpha subunit), the key gene of methanogenesis and methane oxidation 44 , yielded ANME-1 sequences ( Supplementary Fig. 11), indicating methane-oxidizing microbial populations in the currently active Ringvent mound.

Discussion
Ringvent provides the first comprehensively characterized example of an active off-axis hydrothermal vent system in Guaymas Basin, at a spreading age of approx. 1.1 million years 1 . Among currently characterized off-axis Guaymas Basin sites, Ringvent provides the clearest example of a sill-associated hydrothermal site that contrasts with seep-like locations on the northern flanking regions 4 . Similar to the Guaymas Basin spreading center 5,6 , seafloor vent features at Ringvent remain spatially congruent with the underlying sill. The gradually cooling sill at ca. 200 m sediment depth sustains subsurface gas accumulation, hydrothermal circulation and upflow focused along the margins of the buried round sill that reaches the seafloor in a ring pattern. Observations of vent fluids of 20-75 °C associated with sulfur-oxidizing microbial mats and sulfur-oxidizing symbiont-dependent Riftia tube worms indicate channelized flow of hot hydrothermal fluids with temperatures of at least 75 °C, close to the thermal range that has been measured in deep DSDP boreholes, such as Hole 481 in the northern spreading center of Guaymas Basin 45 . Beyond these localized hot spots, the thermal gradients in sediments within and immediately adjacent to the ring structure are almost an order of magnitude steeper than those in the surrounding seafloor sediments ( Supplementary Fig. 2, Table 1). Thus, warm subsurface fluids are likely to impact the entire ring zone (Fig. 7).
In the hydrothermally active southern trough of Guaymas Basin, the mineralogy of active high-temperature and extinct vent structures suggests a paragenetic sequence of mineralization over the life cycle of a vent, with high-temperature deposition of sulfides, followed by moderate-temperature deposition of sulfates, calcite, and barite, and concluding with lower-temperature deposition of amorphous silica 18 . Extinct hydrothermal mounds in southern Guaymas Basin have mineralogical similarities to Ringvent, and analogous silica-dominated mounds and chimneys exist at other hydrothermal sites including the Juan de Fuca Ridge 46 , the Galapagos Ridge 47 , the Central Indian Ridge 48 , and TAG mound in the Trans-Atlantic Geotraverse hydrothermal field 49 . In each of these settings, it is inferred that a silica-rich hydrothermal fluid is conductively cooled, becoming saturated in silica, and variably mixed with seawater followed by precipitation upon further cooling. The temperatures of amorphous silica deposition derived from thermodynamic models and some fluid inclusion studies at analog sites are in the range of ca. 50-150 °C 50 . The consistent recovery of hydrothermal silicates during this and previous sampling at Ringvent 22 indicates hydrothermal silica mobilization and subsequent precipitation as a characteristic process, with the rapidly deposited diatom-rich sediment of Guaymas Basin as the most likely source 2 . The predominance of silicates, and the absence of hydrothermal sulfides from the available samples constrain the thermal regime within the halo of the sill underlying Ringvent. The maximal temperatures of 315 °C that have been documented for hydrothermal fluids at the southern Guaymas Basin spreading center 51 are unlikely to occur at Ringvent presently.
The occurrence of silica nodules in subsurface sediments can provide a window into the history of hydrothermal activity, based on the working hypothesis that these nodules were originally deposited near or at the seawater interface, similar to coastal hot spring sinter deposits in Baja California today 17 . The silica nodule at 2.35 m depth in core P11 indicates a previous event of hydrothermal silica mobilization and deposition in the central area of Ringvent. The potential time horizon is rather broad, with a minimum age of 2350 years, based on an approximate sedimentation rate of 1 mm/y for the Sonora Margin slopes just east of Guaymas Basin, or maximally 8217 years ago based on the local 14 C-derived sedimentation rate for core P11 (Supplementary Data 4). Today, hydrothermal conditions with potential for ongoing silica deposition are found on the ring mound in localized hotspots, sometimes within small depressions and gullies that may have originated during episodes of increased hydrothermal activity. www.nature.com/scientificreports www.nature.com/scientificreports/ A history of variable hydrothermal conditions is also evident from the sedimentary carbonates in core P11. Currently, methane with a distinctive δ 13 C isotopic signature between biogenic seep methane and hydrothermal Guaymas Basin methane permeates the sediment column. The oxidant-replete conditions in the piston-cored sediment column indicate that this methane is not produced in-situ, but appears to have a subsurface source, consistent with a mixed thermogenic/microbial δ 13 C composition (Fig. 7). This methane flux supports methane-oxidizing archaea that form a minor component within the predominantly heterotrophic sedimentary microbial community (Fig. 6), but do not produce enough methane-derived DIC to be visible in δ 13 C-DIC profiles or in contemporary formation of 13 C-depleted carbonates at the sediment surface. Yet, strongly 13 C-depleted, partially methane-derived carbonate nodules appear in the upper sediment column and indicate a relatively recent, strong methane flux, and a larger methane-derived DIC pool. In other words, the 13 C-depleted carbonates in core P11 preserve a paleosignal of methane oxidation to DIC and subsequent incorporation into carbonates that contrasts with minimally δ 13 C-depleted porewater DIC today. By applying a 1 mm/y sedimentation rate from the Guaymas Basin slopes, the depth of the carbonate nodule layer at 1 to 1.20 m translates into an age of 1000 to 1200 years; local 14 C-derived sedimentation rates for core P11 (Supplementary Data 4) yield an age between ca. 4500 to 6700 years. Shallower, methane-imprinted nodules persist in core P11 until ca. 0.5 m depth, and indicate that methane flux and methane-derived carbonate precipitation have slowed down gradually. These timelines imply that carbonate nodules should have formed at former sediment surfaces where sulfate would have been available, and not later within the sediment column.
Today, the concentration and isotope profiles of porewater sulfate, sulfide and DIC show that, of all Guaymas Basin sampling locations, the Ringvent core is most strongly influenced by seawater. In hydrothermal areas, seawater inmixing is commonly connected to hydrothermal circulation via recharge. We hypothesize that evolving hydrothermal circulation has changed the central basin of Ringvent from an active hydrothermal area characterized by silica mobilization and methane venting, to a seawater-influenced recharge zone where seawater is mixed into methane-rich subsurface fluids (Fig. 7). We propose an evolutionary sequence starting with hot, freshly emplaced sills driving intense hydrothermal circulation, silica mobilization at 100-200 °C 18 , followed by precipitation of seafloor silicate minerals, gradually cooling conditions that allow for sulfate-dependent microbial methane oxidation within a thermal limit ca. 80 °C 19 , and deposition of methane-imprinted carbonates, as seen in the sediment column of core P11.
The microbial 16S rRNA gene sequencing results show that the key lineages of microbial methane and sulfur cycling, ANME archaea and deltaproteobacterial sulfate reducers, are less conspicuously present in core P11 than in active hydrothermal sediments of Guaymas Basin, where they often dominate the archaeal and bacterial communities 19,38 . Detecting ANME-1 sequences in core P11 indicates potential for sulfate-dependent methane oxidation, but with the caveat that DNA data do not provide a quantitative proxy for current levels of cellular activity and process rates. In a conservative interpretation, these sequence signatures would indicate a minor component of the sedimentary microbial community, potentially a remnant population of a methane-oxidizing and sulfate-reducing hydrothermal community that is now largely replaced by heterotrophic bacteria and archaea that are common in marine subsurface sediments. At present, the non-sulfidic conditions in the P11 sediment column are likely to attenuate anaerobic methane-oxidizing activity. The conspicuously reduced species richness in core P11 (Fig. 6) suggests that disturbances have purged these microbial communities, for example strong methane seepage or hydrothermal activity that commonly reduce microbial diversity in marine sediments 35,52 . So far, species richness in this core has not fully recovered, and appears impoverished in particular by comparison to core P10, obtained nearby but outside of Ringvent.
Drawing on the available evidence to infer evolutionary scenarios for sill-driven hydrothermal activity, Ringvent provides an instructive example for the hydrothermal stage of an off-axis site transitioning into a cold-seep-like system. It is likely that most sill-hosted systems will transition into cold seeps that persist for a considerable time, since high-permeability conduits will continue to advect heat from deep fluids after the host sill is cool and flow is primarily compaction-driven. It is thus not surprising that seep-like sites are more commonly detectable than shorter-lived hydrothermally active sites 4 . In addition, it is possible that sills intruding at great depth within thick sediments, such as those close to the transform fault that separates Guaymas Basin from the Sonora Margin, cool more slowly due to greater insulation and the inefficiency of hydrothermal flow to those depths. These sites, appearing as cold seeps, may have greater longevity than shallow-sill sites and also mine a larger subsurface methane reservoir. Such cold seep sites in Guaymas Basin carry a signature of their hydrothermal origin: δ 13 C-CH 4 signatures are distinctly heavier than those for biogenic methane on the Sonora Margin, and fall into the mixed thermogenic/biogenic range that is expected for methane in sill-impacted sediments 4 .
This scenario for an evolving Ringvent system should be tested further. For example, subsurface drilling at off-axis sites should uncover geochemical and microbial evidence for hydrothermalism 53 which increasingly resembles the impact of fully developed venting at spreading centers, as depth and proximity to the heat source increase. Deep drilling at Ringvent should constrain the patterns of hydrothermal circulation and recharge, and allow the development of a well-constrained chronology for hydrothermal activity. A deep-sea drilling cruise to Guaymas Basin (IODP Expedition 385) is targeting Ringvent to comprehensively investigate the subsurface foundations for off-axis hydrothermal venting in Guaymas Basin. Porewater chemistry. Porewater was obtained from freshly collected sediments on RV El Puma by centrifuging ca. 40 ml sediment samples in 50 ml conical Falcon tubes for ca. 5 to 10 minutes, using a Centra CL-2 Tabletop centrifuge (Thermo Scientific) at approx. 1000 g, until the sediment had settled and produced ca. 8 to 10 ml of porewater. For porewater sulfate measurements, 1 ml subsamples of the overlying porewater were drawn into syringes and injected through 0.45 μm filters into screwcap Eppendorf vials, each acidified with 50 µl of 50% HCl, and then gently bubbled with nitrogen for 4 min to remove sulfide; the samples were then stored at 4 °C before shipping and analysis. These samples were used for barimetric sulfate quantifications and subsequent determinations of δ 34 S isotopic values at the EaSI lab, University of Texas at El Paso. Barium sulfate (BaSO 4 ) was collected via the addition of BaCl 2 and subsequent centrifugation. The BaSO 4 was transferred into a pre-weighed 2.5 ml sample tube and was twice re-suspended with 2 ml of deionized water followed by centrifugation to remove dissolved salts such as NaCl. The samples were dried, and the weight of the sample was obtained from the difference between tube weight with sample and empty tube weight. For sulfur isotope analysis, ~0. 45  were analyzed using an Elementar ® Pyrocube, connected to a GEOVisION ® isotope ratio mass spectrometer.

Methods
The standard deviation (1σ) for replicate sulfur isotope measurements of standards was 0.1‰. All sulfur isotope values are reported in per mil relative to Vienna Canyon Diablo Troilite (VCDT). For porewater sulfide analysis, 1 ml porewater subsamples were drawn into syringes, filtered immediately through 0.45 μm filters, and placed in Eppendorf sample vials each containing 0.1 ml of 0.1 M zinc acetate solution to preserve the sulfide as zinc sulfide until analyzed. Sulfide was quantified spectrophotometrically at UNC-Chapel Hill using the methylene blue method 54 . Filtered but unamended porewater samples, also stored at 4 °C, were used for quantifying multiple stable ions, including sulfate, by ion chromatography at GEOMAR, Kiel, Germany 55 .
To measure concentrations and δ 13 C signatures of dissolved inorganic carbon (DIC), 2 ml of unamended porewater from each sediment horizon were injected into evacuated serum vials (30 ml) and stored upside down at −20 °C. After the cruise, the samples were acidified with phosphoric acid, and measured by GC-IRMS as described 56 . For combined concentration and δ 13 C analysis of methane, 2 ml sediment subsamples were added to 30 ml serum vials containing 2 ml of 1 M sodium hydroxide solution, sealed with thick butyl rubber stoppers, crimped with aluminum seals and stored at 4 °C. Since cores were retrieved unpressurized, outgassing may have impacted in particular the measurements of methane concentrations near and above saturation, 1.5 mM; however, no gas cavities were observed. After the cruise, the methane samples were analyzed by headspace gas chromatography-flame ionization detection (GC-FID) at Florida State University 57 . Gas samples were analyzed for δ 13 C by injecting 0.1 to 0.5 ml of sample into a gas chromatograph interfaced to a Finnigan MAT Delta S isotope ratio Mass Spectrometer inlet system as previously described 58 . Small amounts of gas were cryo-concentrated before isotopic measurements. Values are reported in the per mil (‰) notation relative to Vienna Pee Dee Belemnite (VPDB). Sediment geochemistry. Samples selected for radiocarbon analyses were freeze-dried, homogenized, and acidified to remove CaCO 3 , allowing for the analysis of remaining organic matter, and preventing the distortion of radiocarbon ages by methane-derived carbonates. Acidification was performed on ~200 mg of sample, which was treated with ~5 ml buffered pH 5 acetic acid solution for ~24 hours to dissolve the CaCO 3 . Samples were then rinsed with Milli-Q water 6 to 8 times to remove the acetic acid. Acidified samples were then freeze-dried again, re-homogenized and stored for 14 C and 13 C analysis. Radiocarbon dating was performed at Lawrence Livermore National Laboratory Center for Accelerator Mass Spectrometry, and a reservoir age of 406 years was (2019) 9:13847 | https://doi.org/10.1038/s41598-019-50200-5 www.nature.com/scientificreports www.nature.com/scientificreports/ used before conversion to calendar years using CALIB REV7.1.0. [http://calib.org]. The δ 13 C values of all the acidified samples used for radiocarbon analyses were approximately −20 to −22‰, as expected for marine primary producers-derived organic matter. Samples selected for carbonate isotope analyses were freeze-dried, homogenized, and roasted under vacuum to eliminate organic matter. Isotope analyses was performed on ~500 µg of sediment using a Kiel devise coupled with a Thermo MAT 253 gas ratio mass spectrometer at the University of California, Santa Cruz. Values are reported in the per mil (‰) notation relative to Vienna Pee Dee Belemnite (VPDB). Reproducibility of in-house standards is 0.07‰ for δ 18 O and 0.03‰ for δ 13 C.
Mineral analyses. Ringvent samples collected at the seafloor by Alvin were studied at UNAM using a transmitted light microscopy (Olympus BX60) coupled with a Motic camera and a low-vacuum Hitachi TM-1000 table-top scanning electron microscope. Stable isotope geochemistry from fibrous aragonite cement samples, and the bivalve shell from Ringvent core P11, were analyzed at UNAM using a mass spectrometer Thermo Finnigan MAT 253 coupled with Gas Bench II, following published guidelines 59 . Bulk mineralogy of seafloor samples was determined via X-ray diffraction (XRD) using an EMPYREAN diffractometer equipped with a fine focus Cu tube, nickel filter, and PIXCell 3D detector operating at 40 mA and 45 kV at UNAM. For this, samples were ground with an agate pestle and mortar to <75 μm and mounted in back-side aluminum holders. The analyses were carried out following previously published procedures 22 . Phase identification was made with PDF-2 and ICSD databases. Mineral phases from core P11 were analyzed separately with a Rigaku Smart Lab XRD using 0.003° resolution, 5°/ minute using a ICDD PDF4+ 2019 database (Ivano Aiello, Moss Landing Marine Laboratory).

Sequence analyses.
To complement taxonomy assignments in VAMPS and to further explore the taxonomic identifications of uncultured groups, archaeal and bacterial sequences were processed with mothur v.1.39.5 62 following the mothur Illumina MiSeq Standard operation procedures 62 . Briefly, forward and reverse reads were merged into 1.6 million contigs (746,180 Archaeal and 831,101 Bacterial) and selected based on primer-specific amplicon length and the following parameters: maximum homopolymer length of 6 and no base ambiguities. Subsequently, 698,385 Archaeal and 763,463 Bacterial high-quality sequences of ca. 330 nucleotides length were aligned against the mothur-recreated Silva SEED v132 database and pre-clustered at 1% dissimilarity. As previously suggested 63 , spurious sequences are mitigated by abundance ranking and merging of rare sequences based on minimum differences of three base pairs. Chimeras were detected and removed using UCHIME de novo mode 64 . Sequences were then clustered, by generating a distance matrix using the average neighbor method, into operational taxonomic units (OTUs, 97% or greater sequence similarity cutoff). OTU classification was performed on mothur using the SILVA v132 database. Community analyses were performed on subsampled datasets (n = 64,890 sequences per sample for Archaea and n = 82,728 sequences per sample for Bacteria). Community structure visualizations and rarefaction analyses were generated using the vegan and phyloseq R-packages 65,66 .
To examine pipeline-derived phylogenetic assignments among methanogenic, methane-oxidizing and sulfate-reducing microbial populations, including hard-to-place uncultured lineages, representative sequences were placed into phylogenetic trees. Pairs of sequenced amplicons were merged by the QIIME function Fastq-join 66 . Representative OTUs were populated using the de novo method in QIIME with a 97% sequence similarity cutoff 67 . After noting the sequence population for every OTU, representative sequences from OTUs containing at least ten sequences were used for downstream phylogenetic analyses, and are listed in Supplementary Materials for reference (Supplementary Data 7). Singletons and chimeric sequences were excluded from analysis using ChimeraSlayer 68 . The Greengenes rRNA database was used for OTU taxonomic assignment 69 . Phylogenetic trees for representative OTUs were inferred with the program package PAUP4.0, using Maximum likelihood distances, transition and transversion rates estimated for each alignment, and Minimum Evolution as optimality criterion 70 . Branching patterns were checked with 1000 bootstrap reruns.
For functional gene analysis, mcrA genes were amplified for ward and reverse primers (5′-GACCAGTTGTGGTTCGGAAC-3′; 5′-ATCTCGAATGGCATTCCCTC-3′) for ANME1 and related archaea 43 . The PCR cycle started with initial denaturation at 95 °C for 1 minute, followed by 30 cycles of denaturation at 95 °C, annealing at 55 °C, and extension at 72 °C of 1 minute each, and concluded by a final 5-minute extension at 72 °C. PCR products were purified using the Wizard SV Gel and PCR Cleanup System (Promega Corporation, Madison, WI, USA) and cloned into plasmid vectors using the TOPO TA Cloning Kit (Life