Evidence for Microbial Carbon and Sulfur Cycling in Deeply Buried Ridge Flank Basalt

See allHide authors and affiliations

Science  15 Mar 2013:
Vol. 339, Issue 6125, pp. 1305-1308
DOI: 10.1126/science.1229240

Under the Sea Floor

Microorganisms living in basaltic sea floor buried beneath sediments derive energy from inorganic components from the host rocks that interact with infiltrating seawater, which brings dissolved oxygen and other trace nutrients with it. Lever et al. (p. 1305) directly sampled the subseafloor community off the eastern flank of the Juan de Fuca Ridge in the Pacific Ocean and found evidence for ongoing microbial sulfate reduction and methanogenesis. Multiyear incubation experiments with samples of host rock confirmed the microbial activities measured in situ.


Sediment-covered basalt on the flanks of mid-ocean ridges constitutes most of Earth's oceanic crust, but the composition and metabolic function of its microbial ecosystem are largely unknown. By drilling into 3.5-million-year-old subseafloor basalt, we demonstrated the presence of methane- and sulfur-cycling microbes on the eastern flank of the Juan de Fuca Ridge. Depth horizons with functional genes indicative of methane-cycling and sulfate-reducing microorganisms are enriched in solid-phase sulfur and total organic carbon, host δ13C- and δ34S-isotopic values with a biological imprint, and show clear signs of microbial activity when incubated in the laboratory. Downcore changes in carbon and sulfur cycling show discrete geochemical intervals with chemoautotrophic δ13C signatures locally attenuated by heterotrophic metabolism.

Subseafloor basaltic crust represents the largest habitable zone by volume on Earth (1). Chemical reactions of basalt with seawater flowing through fractures release energy that may support chemosynthetic communities. Microbes exploiting these reactions are known from basalt exposed at the seafloor, where the oxidation of reduced sulfur (S) and iron (Fe) from basalt with dissolved oxygen and nitrate from seawater supports high microbial biomass and diversity (2, 3). Multiple lines of indirect evidence that include textural alterations (4), depletions in δ34S-pyrite (FeS2) (5) and δ13C-dissolved inorganic carbon (DIC) (6), and DNA sequences from borehole observatories (7, 8) suggest active microbial communities in subseafloor basalt.

We combined sequencing of genes diagnostic of microbial methane and S cycling with geochemical and isotopic analyses of C and S pools and laboratory-based incubations to directly identify microbial ecosystem components in deep subseafloor basalt. The 3.5-million-year-old basement at site U1301 was sampled during Integrated Ocean Drilling Program (IODP) Expedition 301 in 2004 (fig. S1) (9). Site U1301, off the eastern flank of the Juan de Fuca Ridge, is covered by a 265-m-thick sediment layer and lies ~2 km south of ODP Site 1026, which it resembles in temperature profile, lithology, and sediment chemistry (9). Given anticipated poor recovery due to brecciation of the upper basement [265 to 350 m below seafloor (mbsf)], coring was restricted to an interval of pillow basalts and massive lavas (351 to 583 mbsf). Sulfate concentrations (~16 mM) and vein carbonates indicate that basalt fluids are derived from seawater, which enters ~55 km south at Grizzly Bare outcrop and discharges near U1301, at Baby Bare and Mama Bare outcrops (9, 10) (fig. S1B). Yet, the basement at U1301 differs from seafloor-exposed basalt in its uniformly high temperature (~64°C) (9) and lack of fresh photosynthesis-derived organic matter, dissolved oxygen, and nitrate (7, 11). These conditions preclude oxygen- or nitrate-dependent microbial S and Fe oxidation (12) but may enable growth of anaerobes, such as sulfate reducers and methanogens, which use sulfate and DIC as electron acceptors.

We sequenced genes encoding the α subunit of methyl coenzyme M reductase (mcrA), a gene unique to methanogens and anaerobic methane oxidizers (13), and the β subunit of dissimilatory sulfite reductase (dsrB), a gene found in sulfate- and sulfite-reducing microbes (14), to indicate the presence of methane-cycling and sulfate-reducing microbes. We detect mcrA in 5 of the 10 samples and dsrB in 4 of the 6 samples tested (table S1), which suggests that these metabolisms are present in this environment.

The phylogenetic diversity of mcrA genes that we identified is restricted to two groups: the Juan de Fuca Methanogen Group (JdFMG), which falls into an uncultivated cluster within the Methanosarcinales, and anaerobic methane-oxidizing archaea (ANME-1) (Fig. 1A). Close relatives of the JdFMG have been identified from paddy and wetland soil (15, 16) and also have been found in marine habitats, including Juan de Fuca Ridge hydrothermal vent chimneys and seafloor-exposed basalt ~100 km west of U1301 (fig. S2) (17, 18). ANME-1 occur widely in marine sediments and methane seeps and are believed to gain energy from the anaerobic oxidation of methane (AOM) (19). Two distinct ANME phylotypes occur at U1301, one closely related to ANME-1 from methane seeps and another clustering with only one other sequence, from subseafloor sediment (fig. S3). We detected JdFMG in 4 and ANME-1 in 3 out of 10 basalt samples. Two samples contained both groups (table S1).

Fig. 1

Phylogenetic trees of functional genes. (A) McrA sequences from borehole U1301B are in bold magenta. Close relatives based on microarray analyses of JdF Ridge hydrothermal vent chimneys and seafloor basalt are in green (18) and cyan (19), respectively. (B) DsrB sequences from borehole U1301B are in bold magenta and sequences from subseafloor sediment off Peru in cyan (22). Bootstrap support (in %, 1000 replications) is indicated at each branching point.

The phylogenetic diversity of dsrB in these samples is limited to one group, the Juan de Fuca Sulfate Reducing Group (JdFSRG), which falls into Cluster IV, a deeply branching dsrB cluster without cultured members, first reported from hydrothermal sediment (Fig. 1B, fig. S4, and table S1) (20). Remarkably, the only other dsr sequences reported so far from the subseafloor—in sediment of the Peru Margin (21)—also fall into this cluster, which is widespread in shallow marine sediment and terrestrial aquifers.

We studied solid-phase S pools by analyzing acid-volatile sulfide (AVS), chromium-reducible S (CRS), and sulfate-S (SO4-S) as a proxy to redox processes and correlate to microbial metabolisms (5, 22). We found dsrB sequences only in a relatively reduced "intermediate depth interval" (~430 to 520 mbsf, samples 14R to 26R) in samples with AVS as the main S pool in alteration halos (14R-1-11)—the visually conspicuous zone surrounding fractures (fig. S1C)—or in host rock (17R-170, 20R-1-57, and 23R-2-21) (fig. S5 and table S1). Samples from this interval have higher AVS, CRS, and total S (fig. S5 and table S2); contain large pyrite fronts (14R-1-65P, 15R-4-142P) (fig. S5); and have lower δ34S-AVS, -CRS, and -SO4-S, compared with the more oxidized upper (1R to 12R) and lower coring intervals (30R to 36R) (fig. S6 and table S1). Consistent with higher Fe3+/FeTotal ratios, which indicate halos to be more oxidized than host rock (table S1), pyrite is generally absent from halos or veins. Outside the intermediate depth interval, the near absence of pyrite from host rock, and mixed clay-Fe-oxyhydroxide–dominated halos and veins, are further evidence of pervasive oxidative alteration.

We analyzed the δ34S signature of pyrite grains to examine micro- and macroscale variations in microbial S cycling (tables S1 and S3 and Fig. 2). Although variable, the δ34S-pyrite grains [–72.4 to 1.2 per mil (‰)] (table S3) are typically lower than those of AVS (–9.3 to –0.2‰), CRS (–13.7 to 0‰), SO4-S (–6.5 to 0‰), mantle S (0‰) (5), dissolved sulfate in bottom sediments at ODP Site 1026 (+30‰) (23), or seawater (+21‰) (Fig. 2). Locally, the δ34S of pyrite grains reach very negative values (–72‰), consistent with the addition of highly 34S-depleted secondary sulfide to basement rock (22). These low δ34S-pyrite values indicate single-step sulfate reduction (24) or repeated cycles of sulfate reduction and S oxidation (25). The co-occurrence of low δ34S-pyrite, dsrB, and mcrA of ANME-1 in two samples (14R-1-11, 17R-1-70) suggests local coupling between methane and S cycling by sulfate-dependent AOM.

Fig. 2

Macro- and microscale distribution of S-isotopic data. On the left, δ34S-depth profile of pyrite granules, analyzed by laser ablation and secondary ion mass spectrometry (SIMS), and bulk S pools (AVS and CRS). On the right, thin-section micrograph showing individual pyrite granules and their δ34S. The dashed magenta line indicates the sampling depth of the thin section. The dashed black lines mark the intermediate depth interval. Pyrite grains with a sufficient diameter for δ34S determination (10 μm) were limited to this interval. Scale bar, 200 μm.

Depth profiles of total organic carbon (TOC) content, δ13C-TOC, and δ13C-carbonate at U1301B are consistent with functional gene and 34S data (Fig. 3). The TOC content is highest in the intermediate depth interval in cores with mcrA, dsrB, and low δ34S-pyrite (Fig. 3A and table S4). The δ13C-TOC is in the range of dissolved organic C (DOC) in fluids from nearby 1026B and Baby Bare Springs (BBS) (Fig. 3B and table S4) and lower than seawater DOC (–21.1‰) (6). The δ13C-carbonate is higher than δ13C-DIC at Site 1026B or BBS (Fig. 3C and table S5) and overlaps with δ13C-DIC of bottom seawater (–1.4‰) (10).

Fig. 3

Depth-related trends in (A) TOC content, (B) δ13C-TOC, and (C) δ13C-carbonate. Cores with functional gene detection are indicated in (A) and (B). Dashed vertical lines indicate δ13C-DOC (B) and δ13C-DIC (C) values from 1026B and BBS. Because the carbonate content of rock samples used in (A) and (B) was too low for analyses, δ13C from carbonate veins are shown in (C). The reduced intermediate depth interval falls between the dashed horizontal lines. All δ13C are in ‰ versus Vienna Pee Dee belemnite (VPDB).

δ13C-TOC values in the upper coring interval (4R to 5R) and near the bottom (23R to 26R; –34.6 to –32.0‰) are close to δ13C-DOC from nearby BBS (–34.6‰) (Fig. 3B). The absence of O2 and the high 13C-TOC depletion relative to carbonate (~–30 to –35‰) suggest C fixation by the reductive acetyl CoA pathway—an anaerobic pathway found in all methanogens and acetogens and certain sulfate and iron reducers (fig. S7 and tables S6 and S7) (26). The presence of dsrB but not mcrA in these samples suggests that sulfate reducers or other groups, but not methanogens, produce this low δ13C-TOC.

δ13C-TOC at the top (2R) and in the intermediate depth interval (–28.4 to –21.6‰) are close to δ13C-DOC from borehole 1026B (–26.1‰) (Fig. 3B) (6). The 13C depletion relative to carbonate is lower than in the other layers (~–20 to –26‰), but also falls in the range of the reductive acetyl CoA pathway (table S7), and, consistent with mcrA detection, could be affected by autotrophic methanogenesis. In addition, elevated heterotrophic activity is possible, because degradation of chemoautotrophy-derived OC—for example, by AOM, methanogenesis, or fermentation—would lower the δ13C-carbonate and potentially raise the δ13C-TOC. In fact, the lowest δ13C-carbonate values (to –5.1‰) were measured in the intermediate depth interval (18R) (Fig. 3 and table S5), consistent with a locally conspicuous input of IC from the degradation of chemoautotrophy-derived OC. The alternative explanation, enhanced breakdown of photosynthesis-derived OC in the intermediate depth interval, is unlikely given that sediment inclusions are absent (9). Similarly, influx of labile DOC or unaltered DIC from seawater is incompatible with the 7 to 11 thousand year greater DOC age compared with bottom seawater and the 4 to 8‰ decrease in δ13C-DIC along the flowpath from Grizzly Bare outcrop to 1026B and BBS, respectively (6, 10).

To rule out a fossil origin of functional genes and the chemical and isotopic signatures, we incubated pieces from the interior of three rock samples used for functional gene analyses (1R-1-79, 14R-1-11, and 23R-2-21) at 65°C in anoxic, sulfate-rich media containing H2, acetate, methanol, and dimethyl sulfide as energy substrates (table S8). After 2 years, aliquots were transferred to fresh media and incubated for another 5 years using triple-autoclaved basalt pieces as substrata. By then, low concentrations of 13C-depleted methane (–54 to –65‰) had formed, indicating the presence of active methanogenic microorganisms (table S9).

The variability in δ34S-pyrite, δ13C-TOC, and δ13C-carbonate indicates that micro- and macroscale geochemical changes related to mineralogy, fracturing, and/or fluid flow strongly influence microbial activity. These chemical microniches may explain the coexistence of sulfate reducers and methanogens at U1301 and in other igneous habitats, despite higher energy yields of sulfate reduction compared with methanogenesis (27). In addition, some methanogens can survive in the presence of sulfate reducers by consuming noncompetitive methylated substrates (28). Because methanogenic substrate usage follows mcrA phylogeny (28), this explanation is consistent with the ability of a close relative of JdFMG to use methanol (16); it is also consistent with the production of biogenic methane in basalt incubations containing sulfate and methanol (table S9 and fig. S8).

Inorganic electron donors used by sulfate reducers and methanogens—e.g., H2—are likely to derive from serpentinization reactions, whereby Fe(II) minerals—e.g., olivine [(Mg, Fe)2SiO4], which is abundant in several basalt horizons at U1301 (9) (fig. S9 and tables S10 and S11)—are oxidized in abiotic reactions with seawater-derived fluids (1). Organic electron donors (for example, short-chain fatty acids and alcohols) are probably produced by breakdown of autochthonous OC (6, 27, 29) or Fischer-Tropsch–type synthesis (30) (table S10). Targeted investigations of potential carbon and energy sources will provide further insights to micro- and macroscale heterogeneity of microbial C and S cycling and thus contribute to a better understanding of chemoautotrophic ecosystems within Earth's oceanic crust.

Supplementary Materials

Materials and Methods

Figs. S1 to S9

Tables S1 to S13

References (3146)

References and Notes

  1. Acknowledgments: We thank B. Jørgensen, M. Sogin, and the IODP Expedition 301 Scientists for advice and support in this project. Funding was obtained from a Schlanger Ocean Drilling Fellowship, a University of North Carolina Dissertation Completion Fellowship, a Marie-Curie Intra-European Fellowship (255135) (all to M.A.L.); the Danish National Research Foundation and the Max Planck Society (both to B. Jørgensen); Europole Mer (to O.R.); the European Research Council Advanced Grant DARCLIFE (to K.-U.H.); the NASA Astrobiology Institute Subsurface Biospheres (to A.T.); the Japan Society for the Promotion of Science (JSPS) Funding Program for Next Generation World-Leading Researchers (NEXT Program) (to F.I.); and the National Science Foundation (NSF-OCE 0622949 and OCE 1129631 to J.C.A.; OCE-0753126 to S.O. and O. R.; and NSF-ODP 0727175 and NSF-STC for Dark Energy Biosphere Investigations to A.T.). We thank three anonymous reviewers for very helpful comments. The geochemical data are available in the supplementary tables. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. government. The functional gene sequence data are available from the GenBank database (accession numbers GU182109 to GU182110 and JX465656 to JX465658).
View Abstract

Navigate This Article