The Ocean Drilling Program (ODP) is uniquely positioned to sample one of the least known and potentially strangest ecosystems on Earth—the microbial biosphere of deep marine sediments and the oceanic crust. The growing international interest in the study of this subsurface biosphere is driven by a variety of factors, including the role of subsurface microbial activity in Earth's biogeochemical cycles, the possibility of life on other planets, and sheer fascination with the nature of life at the apparent margin of existence.
Nearly 20 years ago, Deep Sea Drilling Project (DSDP) experiments with methane concentration and radiotracer uptake first demonstrated active microbial processes in cores recovered from deeply buried marine sediments (Oremland et al., 1982a; Whelan et al., 1986; Tarafa et al., 1987). Over the last 15 years, studies of ODP cores have extended our understanding of those processes (e.g., Cragg et al., 1992; Getliff et al., 1992) and consistently identified abundant microbes in deeply buried oceanic sediments (e.g., Cragg et al., 1990, 1992; Thierstein and Störrlein, 1991; Parkes et al., 1994, 2000). Microbes have been recovered from burial depths as great as 800 meters below the seafloor (mbsf) (Shipboard Scientific Party, 1999). Recent contamination tracer experiments suggest that the microbes reported from ODP cores are indeed inherent to the drilled sediments (Smith et al., 2000b).
The number and biomass of prokaryotes in the subsurface biosphere of oceanic sediments have been estimated by extrapolation from direct counts of sedimentary microbes at a small number of ODP sites. Based on that extrapolation, the marine subsurface biosphere may compose as much as one-tenth (Parkes et al., 2000) or even one-third (Whitman et al., 1998) of the world's living biomass. In situ metabolic activity by at least a portion of this biosphere is spectacularly demonstrated by hydrates of methane produced by microbes in deep-sea sediments. On a global scale, these hydrates contain 7,500 to 15,000 gigatons of carbon—four to eight times as much carbon as in living surface biosphere and soils combined (Kvenvolden, 1993). Pore water chemical studies (Borowski et al., 1996) and recent microbiological discoveries (Hinrichs et al., 1999; Boetius et al., 2000) suggest that, on an ongoing basis, the CH4 produced in deep-sea sediments is primarily destroyed by the sulfate-reducing activity of microbes in overlying sediments.
Despite these recent advances, very little is known about the nature and activity of life in deep marine sediments. In particular, we know almost nothing about (1) the continuity of subsurface life from one oceanographic region to another; (2) the specialized metabolic properties, if any, that are required to survive in deeply buried marine sediments; or (3) the conditions under which subsurface microbes are active or inactive and living or dead.
There is abundant evidence of both microbial populations and microbial activity in subsurface marine sediments throughout the world ocean. Prokaryotic cells have been found in surprisingly high numbers in buried sediments at every site that has been assayed for their presence (Parkes et al., 2000). The abundance of those cells varies in a systematic and fairly predictable manner. For example, deeply buried shelf sediments from the Peru margin (high surface-ocean productivity and shallow water depth) contain 108–109 cells/cm3, and sediments from the eastern equatorial Pacific (low surface-ocean productivity and abyssal water depth) contain only 106 cells/cm3 (Parkes et al., 2000).
Pore water chemical data from hundreds of DSDP and ODP sites document the occurrence of subsurface catabolic activity in sediments throughout most of the deep ocean (D'Hondt et al., 2002). Microbial sulfate reduction, methane production, and methanotrophy are common processes in deeply buried marine sediments. Other catabolic processes are known to occur in subsurface marine sediments but have been studied very little (such as manganese and iron reduction).
Despite the ubiquity of microbial cells in deeply buried marine sediments and the clear pore water evidence of in situ microbial catabolism, the identity and structure of these communities and the metabolic adaptations of the microbes that constitute them remain largely unknown. Most probable number (MPN) experiments have demonstrated that viable cells are present in deeply buried marine sediments (Parkes et al., 2000). However, these viable cells represent only the barest fraction (0.00001% to 0.6%) of the total cells enumerated in the sediments sampled (Parkes et al., 2000). The extent to which this discrepancy between enumerated and viable cells reflects a culturing bias, known also from surface sediments, or the extent to which it reflects a real difference between a small active population and a very large inactive (dormant or dead) population remains to be determined.
The importance of this issue for our understanding of subsurface population structure and metabolic adaptation is underscored by estimates of the mean sulfate reduction per enumerated subsurface cell (D'Hondt et al., 2002). If all of these enumerated cells are alive, their rates of SO42– reduction are zero to two orders of magnitude lower in the ocean-margin anaerobic methanotrophy zone and four or more orders of magnitude lower in open-ocean sediments than per-cell rates inferred in surface marine sediments (Jørgensen, 1978; Knoblauch et al., 1999; Ravenschlag et al., 2000; D'Hondt et al., 2002). In contrast, if subsurface cells actually utilize SO42– at the lowest rates inferred for cells in surface marine sediments, as few as 1 in 100 may be actively respiring in the sulfate-reducing methanotrophy zone of the most microbially active sites and fewer than 1 in 10,000 is actively respiring at the most microbially active open-ocean sites. In short, most of the subsurface microbes enumerated by direct microscopy in marine sediments must be either adapted for extraordinarily low levels of metabolic activity or dormant—or even dead. This conclusion is supported by available estimates of mean generation times of up to 1 m.y. for deep subsurface microbes (Parkes et al., 2000).
The metabolic diversity and rates of microbial processes in deep subsurface sediments can be inferred from a broad range of geochemical information, including modeling of pore water profiles of ions, gases, and low molecular weight organic molecules, mass balance calculations of changes in solid phase constituents, and stable isotope fractionation. Basically, the same types and sequences of microbial processes appear to occur deep in the seafloor as are known from the much more active surface sediments of ocean margins. The mechanisms and regulation of the exceedingly slow hydrolytic degradation of macromolecular organic compounds are, however, only poorly understood. So, too, are the fermentative pathways that produce substrates for the terminal mineralization processes such as sulfate reduction or methanogenesis.
The buildup of bicarbonate and ammonium are indicators of the diagenesis of organic material in all marine sediments. Sulfate reduction dominates down to the depth of sulfate depletion, many tens or hundreds of meters below the seafloor, where it is followed by methanogenesis. Although methane formation is very slow, the continuous production of a diffusible gas over millions of years results in vast methane accumulations, either dissolved in the pore water or in the condensed form of gas hydrates. In both cases, an upward flux of methane reaches the sulfate zone and supports an interface of enhanced microbial activity based on methane oxidation. In shallow marine sediments, this anaerobic process is catalyzed by a syntrophic community consisting of archaea, which convert methane back to an intermediate such as hydrogen or acetate, and sulfate reducing bacteria, which oxidize the intermediate (e.g., Hoehler et al., 1994; Valentine and Reeburgh, 2000; Boetius et al., 2000; Orphan et al., 2001). Based on pore water modeling of sulfate and methane profiles, the same process appears to drive a significant part of sulfate reduction in the seabed (e.g., Borowski et al., 1999; D'Hondt et al., 2002). This interface is of particular importance, since it constitutes a barrier against the escape of methane up into the ocean water and eventually into the atmosphere.
In deep-sea sediments, such as the Peru Basin sites drilled during Leg 201, manganese oxide may provide an important oxidant of organic material, and its reduction can be traced tens of meters into the seafloor (Yeats, Hart, et al., 1976). The reduction of iron oxides expectedly plays a greater quantitative role, and iron(III) bound in mineral phases may provide a slow but continuous source of oxidation potential over 105–106 yr (Raiswell and Canfield, 1996).
H2 is an important intermediate in the microbial degradation pathways of ocean-margin sediments, and its pore water concentration is strictly regulated by the uptake potential of the microbial community and the energy yield of their H2 metabolism. Thus, the H2 partial pressure in the sulfate reduction zone is maintained below the threshold level required by archaea to drive methanogenesis (Hoehler et al., 1998). The key role of H2, known from the metabolic competition among microbial populations in surface sediments, may also be critical for the deep subsurface biosphere. The potential sources for microbial energy metabolism need to be surveyed with an open mind toward new and unexpected types of redox processes and mineral surface reactions.
The phylogenetic and physiological diversity of deep sediment communities remains virtually unknown. Only two isolates of sulfate-reducing bacteria from subsurface sediments (80 and 500 mbsf) have been characterized (Bale et al., 1997). These isolates (from a single site in the Japan Sea) are of a new barophilic species, Desulfovibrio profundus. Its unusually wide growth temperature range (15°–65°C) has no counterpart in any other known sulfate-reducing bacterium. It is metabolically flexible; it possesses the capability to reduce oxidized sulfur species, nitrate, and Fe(III). Whether deeply buried sulfate-reducing communities throughout the world ocean are dominated by close relatives of D. profundus or are composed of a host of other organisms remains to be tested.
The record of methanogenic isolates from the subsurface is surprisingly spare. MPN enumerations of methanogens in deep marine sediments have yielded cultured methanogens in much smaller numbers than sulfate-reducing bacteria (in the Peru margin; Cragg et al., 1990) or not at all (in the Japan Sea; Cragg et al., 1992). These surveys have, to our knowledge, not led to the description of new methanogen species from the marine subsurface. Hence, the phylogenetic composition of marine subsurface methanogenic communities remains essentially unknown.
Organisms responsible for methanotrophy in nearshore surface sediments have been biomarker fingerprinted and phylogenetically identified but not yet cultured (Hinrichs et al., 1999; Boetius et al., 2000; Teske et al., 2002; Lanoil et al., 2001). Whether or not similar microbial communities (composed of sulfate-reducing bacteria and previously unknown members of the archeal Methanosarcinales) are responsible for methanotrophy in the more deeply buried biosphere throughout the world ocean (in both ocean-margin and open-ocean environments) remains to be seen.
Novel forms of bacterial metabolism with subsurface potential are constantly being discovered. For example, systematic studies of sulfate- and sulfur-reducing bacteria and archaea have shown that many representatives of these organisms, among them an astonishing set of phylogenetically very deep lineages (Vargas et al. 1998), share an unexpected capacity for Fe(III) reduction (Lonergan et al. 1996). A Thermus sp. isolated from a deep South African gold mine used O2, NO3–, Fe(III), S0, Mn(IV), Co(III), Cr(VI), and U(VI) as electron acceptors (Kieft et al. 1998). Respiration of metal oxides could allow bacteria and archaea a respiratory mode of life even after other electron acceptors, including oxidized sulfur species, become depleted with increasing distance from the oxidized biosphere. The metabolic flexibility of D. profundus and the South African Thermus sp. allows multiple scenarios of subsurface phylogenetic diversity. One possible scenario is that a certain microbial community becomes buried below the sediment surface and basically persists in its phylogenetic diversity and physiological potential over millions of years. This community may be responsible for NO3– reduction, manganese reduction, iron reduction, and SO42– reduction throughout the vertical expanse of a single sediment column—and even dominate the subsurface respiratory realm throughout the sediments of the world ocean (at least within a broad temperature range, such as 0° to 30°C or 30° to 60°C). Such a persistence was suggested to explain the hyperthermophilic archaeal rDNA (ribosomal deoxyribonucleic acid) sequences in subsurface sediments of the West Philippine Basin, supposedly representing microbial relics originating from past submarine hydrothermal activities >2 m.y. ago (Inagaki et al., 2001). A second (and perhaps more likely) scenario is that the phylogenetic composition of subsurface communities may be shaped by variables other than the type of electron acceptors available. For example, it may be controlled by electron donor availability, micronutrient availability, or the ability of well-tuned species or communities to out-compete each other under slightly different local environmental conditions, such as different concentrations of metabolic products and reactants. Also, physical factors such as available pore space, ability to migrate in the sediment, interactions with mineral surfaces, or distance to solid substrates may be important.
The subsurface biosphere of marine sediments may affect the surface Earth in a variety of ways. It is now widely recognized that release of CH4 from marine sediments may affect atmospheric carbon stocks and climate (Dickens et al., 1995; Dickens, 2000; Kennett et al., 2000; Hesselbo et al., 2000; Hinrichs, 2001). It is less widely recognized that SO42– reduction by the buried biosphere may also change Earth's surface chemistry and climate. Such reduction is a major sink of SO42– from the world ocean (Holland, 1984). Because SO42– is the second most abundant anion in seawater (Pilson, 1998) and SO42– reduction followed by sulfide precipitation results in the production of two equivalents of alkalinity per mole, subsurface SO42– reduction may affect total oceanic alkalinity and, consequently, the partitioning of CO2 between atmosphere and ocean over geologic time (D'Hondt et al., 2002). The ultimate effect of this process on the surface Earth will depend on the extent to which reduced sulfur is fixed in the sediment rather than diffusing back into the overlying ocean to be oxidized. Furthermore, NO3– reduction by the subsurface biosphere may be a net sink of biologically accessible nitrogen from the world ocean (Moore, Taira, Klaus, et al., 2001).