Well-hidden methanogenesis in deep, organic-rich sediments of Guaymas Basin

Deep marine sediments (>1mbsf) harbor ~26% of microbial biomass and are the largest reservoir of methane on Earth. Yet, the deep subsurface biosphere and controls on its contribution to methane production remain underexplored. Here, we use a multidisciplinary approach to examine methanogenesis in sediments (down to 295 mbsf) from sites with varying degrees of thermal alteration (none, past, current) at Guaymas Basin (Gulf of California) for the first time. Traditional (13C/12C and D/H) and multiply substituted (13CH3D and 12CH2D2) methane isotope measurements reveal significant proportions of microbial methane at all sites, with the largest signal at the site with past alteration. With depth, relative microbial methane decreases at differing rates between sites. Gibbs energy calculations confirm methanogenesis is exergonic in Guaymas sediments, with methylotrophic pathways consistently yielding more energy than the canonical hydrogenotrophic and acetoclastic pathways. Yet, metagenomic sequencing and cultivation attempts indicate that methanogens are present in low abundance. We find only one methyl-coenzyme M (mcrA) sequence within the entire sequencing dataset. Also, we identify a wide diversity of methyltransferases (mtaB, mttB), but only a few sequences phylogenetically cluster with methylotrophic methanogens. Our results suggest that the microbial methane in the Guaymas subsurface was produced over geologic time by relatively small methanogen populations, which have been variably influenced by thermal sediment alteration. Higher resolution metagenomic sampling may clarify the modern methanogen community. This study highlights the importance of using a multidisciplinary approach to capture microbial influences in dynamic, deep subsurface settings like Guaymas Basin.

sill at ~350mbsf.The third site, U1547, is located ~20km NW of the spreading center and is experiencing active thermogenic alteration via a shallow (~75mbsf), young sill.U1547 is located inside of the circular hydrothermal mount known as Ringvent (14,15).While located in a different location than U1546, U1547 may be indicative of conditions at U1546 when its sill was first emplaced.
Using shipboard measurements of porewater geochemistry and temperature from hole B, we identified distinct physicochemical patterns down-column at each site (Figure 2), including different temperature gradients and SMTZ depths, suggestive of variable microbial activity (16).
The temperature gradient at U1547 (511°C/km) is nearly double that at U1545 and U1546 (221 and 225°C/km, respectively).The similar gradients at U1545 and U1546 implies the sill at U1546 had reached thermal equilibrium with the surrounding sediment (14).All three temperature gradients are significantly higher than the global average (25°C/km).The SMTZs are at ~45 mbsf at U1545 and at ~110mbsf at U1546 and U1547 (both of which have magmatic sills present).At all three sites, the methane levels below the SMTZ are as high as 2.5-3mM, but at U1546 and U1547 two additional peaks are present at greater depths.At U1546, methane concentrations reach 6.09 mM at 237.6 mbsf and 3.4 mM at ~332.3 mbsf.The methane levels at U1547 are 1-2 orders of magnitude higher, reaching 60.4 mM at 131.9 mbsf and 125.7 mM at 151.2 mbsf.We build this study on these preliminary findings.

Metagenomic binning
The scaffolds from the MEGAHIT assemblies ( >2000 bp) were binned using MetaBAT v2.12.1 (Kang et al., 2015) and CONCOCT v1.1.0(Alneberg et al., 2014), and resulting bins were assessed and filtered using DAS Tool v1.1.2 (Sieber et al., 2017).The accuracy of all the bins was evaluated by calculating the percentage of completeness and gene duplication using CheckM lineage_wf v1.0.5 (Parks et al., 2015).Bins which were >50% complete but > 10% redundant were manually refined using Anvi'o's anvi-refine function.Ultimately, only bins >50% complete and <10% redundant were kept in the dataset.Bins assembled from scaffolds which map to the drilling fluid and blank samples were used to decontaminate the overall bin dataset.The finalized bins were surveyed for mcrA sequences found in the general assembly that may have mapped back to them.
We did not find this to be the case.As such, the bins are not discussed or reported in this study.

Isotopic Analysis
Freshly preserved sediment samples U1545B_8H_3, U1546B_29F_2, U1546B_49F_3, U1546B_54F_2, U1547B_9H_2, U1547B_21F_2, and U1547B_24F_2 were prepped for multiple isotopologue measurements on the IRMS Panorama at the University of Maryland Panorama Laboratory as follows.Prior to purification, methane concentration in head space gas was measured using a Shimadzu GC8A with FID.Methane from samples with high nitrogen concentrations was preconcentrated by freezing the sample into a liquid nitrogen-cooled U-trap packed with 100-120 mesh particle size HayeSep D porous polymer and then warming the trap to −115C using an ethanol slush to release most N2.The preconcentrated methane plus nitrogen was then transferred to the injection loop filled with silica gel and liquid nitrogen-cooled of a SRI 310C gas chromatograph, from which it was injected into a 1/8" diameter 25 Ft mol sieve 5 Å column at 54C in a He carrier.The methane peak was detected using a TCD and then trapped in a liquid nitrogen-cooled collection loop filled with silica gel.The collector loop was isolated and residual helium was pumped away while the loop was kept at liquid nitrogen temperature, which retained only the purified methane.The methane yield was determined monometrically, and the methane was transferred to a liquid-nitrogen cooled sample tube containing silica gel.After purification, methane samples were analyzed using the Panorama high resolution mass spectrometer (Nu Instrument) at The University of Maryland -College Park.
Methane was measured using sample standard bracketing relative to a working gas (UMD-1) and ratio data are reported in per mil using the δ and Δ notations, respectively: and where i R or j R represent the isotope ratio by number i, including 13 C/ 12 C and D/H, or isotopologue ratio j, including 13 CH3D/ 12 CH4 or 12 CH2D2/ 12 CH4, standards are Standard Mean Ocean Water (SMOW) for D/H and Pee Dee Belemnite (PDB) for 13 C/ 12 C, and stochastic refers to the stochastic isotopologue ratio for the sample.

DISCUSSION
To partition and compare the percent microbial methane (PMM) throughout the sediment columns, we constructed isotopic mixing models that take into consideration δ 13 C, δD, Δ 13 CH3D and Δ 12 CH2D2 (17,18).The traditional isotope and hydrocarbon values of the methane sample from ~20m below the SMTZ of U1545 is within the well-constrained microbial fields (see Figure 4a-b) (19)(20)(21)(22).Thus, for the microbial endmember, we chose δ 13 C and δD values that are slightly more negative (δ 13 C = -80 vs. -76‰; δD =-225 vs -200‰).Energy-rich environments where methane is rapidly produced, such as axenic microbial cultures and natural samples from boreal lake wetlands, yield disequilibrated methane isotopologue compositions (Δ 13 CH3D = -4 to 4, Δ 12 CH2D2 = -60 to -15), credited to kinetic effects (23)(24)(25)(26).Methane isotopologues can only be used as geothermometers at equilibrium, thus temperatures cannot be assigned at disequilibrated values.However, in energy-starved environments, such as the marine subsurface, methanogenesis and anaerobic oxidation of methane (AOM) may produce or cycle methane towards near equilibrium isotopologue values.This is suggested to be a result of high enzymatic reversibility caused by limited substrate and energy availability (27)(28)(29)(30)(31)(32).Notice that while methane samples that fall within the 'microbial' fields of U1545 and U1547 have the expected Non-microbial methane can be of thermogenic or abiotic origin.Hence, we created two mixing models: with microbial and thermogenic endmembers; with microbial and abiotic endmembers (see Supplemental Figure 1-3).Thermogenic methane can yield either equilibrium or disequilibrium signatures, depending on whether it is sourced from mature or immature, respectively, hydrocarbon precurors (26,33).We know pyrolysis products in Guaymas sediments are thermally mature, and laboratory studies have assigned δ 13 C values of -20 to -23‰ (8,34,35).